CA2095442C - Downward compatible agv system and methods - Google Patents

Abstract

An automated guided vehicle (AGV) control system which is downward compatible with existing guidewire systems pro-viding both guidewire navigation and communication and autonomous navigation and guidance and wireless communication bet-tween a central controller and each vehicle. Autonomous vehicle navigation comprises travel over paths marked by update mark-ers which may be spaced well apart, such as fifty feet. Redundant measurement capability comprising inputs from linear travel encoders from the vehicle's drive wheels, position measurements from the update markers, and bearing measurements from a novel angular rate sensing apparatus, in combination with the use of a Kalman filter, allows correction for navigation and guid-ance errors caused by such factors as angular rate sensor drift, wear, temperature changes, aging, and early miscalibration during vehicle operation. The control system comprises high frequency two-way data transmission and reception capability over the guidewires and via wireless communications. The same data rates and message formats are used in both communications sys-tems.

Description

2 0 9 5 4 ~ 2 PCr/US9l/08892 DOWNWA7.~ CO~qpA~TRT 7;~ AGV SYSTEM AND ME~IODS
Field of Invention The f ield of the invention is control, ,_ ; r~7tion systems, and automatic navigation and q~7j~7An~-e of vehicles, including vehicles that navigate without a driver on board either by self-contained navigation and guidance with occasional calibrating updates or, alternatively, by following a guide wire which is activated by an AGV controller which is not on the vehicle or, otherwise, activated by energy sources on the vehicle itself.
Backaro~nf7 And Descri~?tion of Related Ar~
Automated guidance over a wire-guide path has been used in the guidance of a driverless automatically --controlled vehicle (AGV) along a desired course have been set forth in U. S . Patent Numbers 3, 009, 525 and 3, 147, 817 issued to Robert DeLiban. In such disclosures, the AGV
followed a traffic path defined by a conductor energized by source not on the vehicle.
- Later U.S. Patents 4,491,570 and 4,902,948 of this assignee describe communication systems and methods for controlling a plurality of task-performing AGV' 5 along a network of guide wires. U.S. Patents 4,791,570 and 4, 902, 948 describe a guide wire logic and co7Dmunications capability which provides for infinite expansion as to the number of guide wire loops and vehicles which comprise the system; a~ 'AteS polling of vehicles of the system not at predetermined times but only upon the occurrence of certain events, causes high data transmission rates to occur over low frequency carriers using the guide wires.
- Today, there are a large number of installations Qf AGV systems which employ guide wires.
7~owever, the cost of installing and remode1ing guide wire --paths has proved to be a deterrent to purchase of new systems and the expansion of older installations.
~ , --2- ao~5442 Factory layout flexibility and related illal " ~ and U~ldliUlldl coât reduction, not realized with guidewire âystemâ, is possible with autonomousiy operating AGV's. Apparatus and methods of gl,;d~ u~.a control of AGV's are found in U.S. Patents 4,908,557 and 4,847,769 and in published European Patent ~i." Li",- 193,985.
U.S. Patent 4,908,557 issued to Masahiro Sudare discloses a control apparatus which navigates along a path defined by update magnets arranged in thefloor such that a null or bipolar signal is prûduced in each detecting sensor of a Hall sensor array located in each AGV. An algorithm is described which calculates ,0 positlon of the magnet based upon treating each sensor as a point or unit of measurement and performs calculations based uporl a minimum distance in units of sensor positions. The precision of measurement is aLdLiaLil,.~ !y d~ -d~.lL upon the physica Ml;.,,JIdCe~ L of each of the Hall sensors and the aLtl~pl~eas of the signal about the transition between the magnet's north and south fields. As such, 15 the precision of measurement of a magnet's position by the method described by Masahiro Sudare is of the order of magnitude of the center-to-center spacing of the Hall sensors. This level of measurement precision produces errors in vehicle bearing estimates which markedly restricts allowable distance of S~yc"dLiù,~
Detween the update magnets in the vehicle path, s;~lliri~dllL precision being ~0 required to provide assurance the vehicle will retain sufficient bearing accuracy to acquire to stay on a planned path between widely separated magnets.
U.S. Patent 4,955,447 issued to I lelall "u~u discloses a guiding method for an AGV c~ JIiaillg guiding by means of a guideline (guidewire) when the AGV is over the guideline and a radio wave system culll~Jliailla~ wireless eq~ i~ "~"~L on 25 board the vehicle and stationary âubstations when the AGV is not over the guideline.
U.S. Patent 4,847,769 issued to Peter J. Reeve discloses a navigation system which carries out a dead reckoning calculation of the vehicle's position based upon inputs from linear and angular measurements from a steering wheel 30 and a bearing and/or a range to a target. The bearing and/or range to the target, dt,lt"",;,lsd by laser bearing finding eq~i",~",L, provides updating data which are .

-3- ~0 ~54S2 used to ~ to peri~Ji~,_'!y reduce errors due to drift and other factors in the heading angle and spatial position of the AGV, angular drift in the steering angie, crabbing angle, and variations in the measured radius of the steering wheel. A
Kalman filter is used to calculate corrective ~ iOl~S which are derived from the5 bearing and/or range to the target measurements. The laser bearing finding equipment cu, "yl isas a laser emitter located at an obstruction free position on the AGV such that the vehicle may confirm its position by seeking a number of targets distributed about the AGV in a factor,v frame of reference.
European Patent ~ iu,- 193,985 describes a grid-wireless system for 0 ~ a free ranging vehicle. The system employs a srid of marker elements which are closely spaced to eliminate the measurement problems encountered with prior known navigation systems.
However, none of the related art adJ, ~ s problems related to c~" ,, ' ' with existing guidewire systems, providing the capability to operate along an 15 existing guidewire path and in an autonomous mode as well. Further, problems re~ated to ",;.,;I"i~;"g the numbers of floor markers required and, therefore, long distance autonomous operation between floor markers and providing ~" " c ~11 iul~,d top loading surfaces for vehicles have likewise not been add~ " "ed in the knownrelated art.

4 20 ~5442 .
RRTR~ SI~ARY AND OB~ECTS OF THE INVFI~TION
An invention in which a guided vehicle follows passive conductors on the floor ib debCribed in ~J.S. Patent 4,613,804, issued 9/23/86, entitled "Floor Position Sensing Apparatus and Method, n invented by R . R . Swenson .
In brief summary, the present invention comprises an automated guided vehicle (AGV) control system which is downward compatible with existing guide-20 wire systems, providing both AGV guide-wire navigation and communication and autnn~ _q navigation and guidance and wireless communication within the same vehicle. The AGV control system comprises an AGV controller, a plurality of types of guide path marking apparatus, at 25 least one AGV capable of a plurality of navigation and guidance modes, including aut~- ~ operation, ~and a two-way communication system between the AGV controller and each AGV.
The AGV controller controls the movement o~
30 each individual AGV along pr~d~t~rmin~ path segments.
Two-way ~ ; -~tions comprise guide-wire carried messages and wireless messages provided over short access period links between the each AGV and the AGV controller Guide path marking types of apparatus comprise vehicle 35 powered and non-vehicle powered guide-wire loops and guide path update markers.
The navigation nd guidanc~ system comp~ ises ,.

WO 92/09941 2 0 9 5 4 4 2 PCr/US91/08892 redundant mea:iuL. L cAr~hi 1 ~ ty 6uch that meaDuL~ - L
errors caused by factors comprising drift t~ _ ~LULe change, wear, and aging are dynAmicAlly evaluated during each AGV operating mission and sensor inputs are

5 calibrated in real time to reduce the effect of such errors . A Kalman f ilter is used to determine calibrating updates. Sensor meaDuL- L precision and calibration and novel navigation methods provide autnn~ - operation such that an AGV operating in the autnn~ - mode 10 experiences an error having a deviation standard no greater than two inches when traveling between update markers which are fifty feet apart. The AGV controller sends and each individual AGV receives and acknowledges each next path sejment end position and exit bearing from 15 that path segment. Thereby, in-;L Lal control of each vehicle over ~LaS~, - Led portions of a path is provided by the AGV controller. From the position and bearing received from the A,V controller, each AGV calculates a non-linear path for self-contained guidance control of 20 the AGV over the path segment whereby vehicle guidance accuracy is i uv~d.
Such vehicle navigation and guidance accuracy provides a system which comprises widely spaced update markers and a resultingly low guide path installation and z5 ~ Al inAj cost. rAurther, passive guide wire apparatus, which i5 manually portable, provides a temporary path for vehicles during r~ - "Al i n~ and the like and a high accuracy guide path for positioning an AGV at a tPrm;nAl.
Through the use of mea~url Ls which provide redundant 3 0 estimates of vehicle distance and bearing and, thereby, COn~;ULLe:~lL estimates of mea~ul L errors, such as angular rate sensor drift, a low cost rate sensor is effectively used in the vehicle navigation and guidance system. Aperiodic sampling of the angular rate sensor 35 apparatus wlth drift corrections, provides an effective redundant meaz~UL~ t in the vehicle sensor matrix. All of the AGV sensors and c i ~Antions systems reside .

WO 92/09941 2 ~ 9 5 4 4 2 PCl/US91/OX892 1 below the top surface of the AGV, freeing the top surface for loading and unloading cargo and for attaching to other vehicles.
Accordingly, it is a primary ob; ect to provide an AGV control system which controls at least one AGV
which navigates over path6 marked by guide-wires and, alternatively, over paths marked by intermittently placed update markers.
It is a further primary object to provide an AGV control system which comprises auto~ cly operating AGV ' s and wireless ; cAtions and is downwardly compatible with existing AGV guide-wire installations.
It is a still further primary object to provide a system for controlling a plurality of, nn~rl, task-performing vehicles whereby the travel paths and tasks performed by the vehicles are strictly controlled by an AGV controller on a path segment by path segment basis.
It is another primary object to provide at least two navigation and guidance systems within at least one AGV, at least one navigation and guidance system providing greater accuracy and precision than at least one of the other navigation and guidance systems, whereby at least one AGV is selectively more accurately and precisely guided over selected segments of the guide path and less accurately and precisely guided over other segments whereby a cost effective selection of guide path markers may be made, based upon accuracy and precision requirements of each guide path segment.
It is another primary obj ect to provide a 3 0 digital computer based, automated guided vehicle controller which provide centralized plAnn i n~ and control for at least one AGV whereby each AGV is sent a control message which defines a limited activity to be performed by the vehicle.
It is yet another primary object to provide an AGV controller which sends position and bearing of the end of a next-to-be-traveled path segment by which the WO 92/09941 2 ~ 9 5 gl ~ 2 PCr/US91/08892 AGV calculates a guide path.
It is another object to provide an AGV
controller comprising a compiler which provides position and bearing, f or the end of each path segment sent to a vehicle, calculated from previously entered input data which def ines a plurality of markers and paths within a factory frame of reference.
It is another object to provide an AGV
controller which i6 pL~yL hly rh:nqe~h~ ~ using a higher plOyL ; n~ language, such as a "C" compiler.
It is another object to provide an AGV control system which comprises update markers in the floor of an AGV path comprising marker to marker spacing which may be widely separated, such as fifty feet between markers, thereby reducing installation and L ~ ; n~ costs .
It is another object to provide an AGV control system which comprises a combination of markers in the floor, the combination comprising update markers and guide-wires .
It is another object to provide an AGV control system which comprises a combination of markers in and on the floor, the combination comprising update markers and guide-wires .
It is another obj ect to provide an AGV control system wherein the update markers in the floor of an AGV
path comprise magnets.
It is another object to provide an AGV control system which comprises update markers in the f loor of an AGV path, the markers comprising magnets oriented such that only a South or a North field is sensible by superiorly ~1; cposPd 6ensors.
It is another object to provide an AGV control system which comprises a combination of markers in and on the floor, the combination comprising guide-wires which are activated by an energy source on an AGV and guide-wires which are activated by an energy source not on the AGV .

~095~ 1~ 8 ~
It is another object to provide an AGV control system which comprises a combination of marker6 in and on the floor, the combination comprising markers which are activated by an energy source on an AGV and markers which 5 are activated by an energy source not on the AGV.
It is a principal object to provide an AGV and AGV controller communication system which assigns tasks by polling each AGV.
It is another principal object to provide0 AGV/AGV controller c ; rations which comprise wireless i rations apparatu8 .
It is another principal object to provide receiving AGV c ; ration5 which acquire an i nt i n~
message in less than five m;11;cPrQn~lc.
It is another principal object to provide AGV
communications which alternatively use wireles6 or guide-wire l;~rntions links.
It is a further principal object to provide an AGV and AGV controller -- I;r-ation system which 2 0 communicates over wireless ~ ; c~tion5 wherein each AGV is polled, not at pr~et~rm;necl time~, but only upon the oc~-lrLt ..ce of certain events.
It is a still further principal object to provide an AGV and AGV controller _ ; rntion system 25 which t ;r-ates over wireless communications wherein each message is acquired by the receiving apparatus in less than five m; 1 1; cer~-n~c.
It is another principal object to provide an AGV and AGV controller ;r-ation system which 3 0 communicates over wireless ~ ; cation5 wherein each message is uniquely addressable to a single AGV.
It is a main object to provide a navigation and guidance system for an AGV which is totally contained below the top surface of the AGV, such that the top 35 surface is free for loading and 11n1t~ ;n~ and for atta~ L to other apparatus.
It is another main object to provide a WO 92/09941 2 ~ 3 ~ PCr/US91/08892 navigation and guidance system for an AGV which comprises two-way . ic~tions ~ aLaLus~ whereby at least task direction is received from an AGV controller and me6sage acknowl ? ', ~ and AGV status is transmitted to the AGV
controller.
It is further main object to provide a navigation and guidance system which comprises a r~AllnA~nry in sensor - _L- L c~r~hi ] ity such that sensor-based Qrrors due to factors comprising t~ aLuLe, wear, drift, aging, and earlier incorrect calibration are quantif iable .
It is a still further main object to provide a navigation and guidance system which recalibrates SenSQr inputs in real time with quantified values derived from proc~sin~ redundant data from the sensors whereby the accuracy and precision of the navigation and guidance system is improved.
It i5 another further main object to provide a Kalman f ilter by which the sensor based error6 are quantified.
It is another object to provide navigation and guidance apparatus for enabling an AGV to ascertain and control its position rather precisely at a predetPrmi neA
area on the floor, such as at a t~rmin~l.
It is another object to provide navigation and guidance apparatus which enables an AGV to ascertain not only its position relative to a floor reference system, but also its heading, by sensing the lateral positions of two sensors on the vehicle that are spaced apart 3 0 longitllA i n~ 1 1 y It is another object to provide position-sensing navigation and guidance apparatus in which only passive elements are required on (o~ in) the floor and all energy required for the sensing of position comes from the vehicle, at least at certain areas such as in a tèrminal .
It is another object to provide navigation and WO 92~09941 PCr/US91/08892 2~9~44~ lo O
guidance position-sensing apparatus in which the passive elements of equipment at the f loor comprise one or more passive loops of electrical conr117t-tt~r.
It is another object to provide a navigation 5 and guidance position-sensing apparatus having a magnetic-signal receiving system that _ tes for undesired signals, such as those received directly from lts transmitting antennas on the vehicle, and responds only to signals received indirectly via floor-mounted lO passive loops.
It is another object to provide a navigation and guidance position-sensing apparatus on an AGV in which two receiving coils are spaced apart on only one high pf --hi 1 ity magnetic core to improve the linearity 15 of response of the signals as a function of the amount of their offset from a passive loop on the floor.
It is another object to provide navigation and guidance app2ratus to enable an AGV to utilize equipment in common to ascertain and control bo1:h its lateral and 20 longitudinal positions relative to a known reference on the floor at, for example, a t~rm;nAl.
It is another object to provide a navigation and guidance system for positioning an AGV in which the AGV is automatically guided to a predetermined station or 25 t~rmin~l by one type of ~~ nre mode and is precisely positioned within the station by another type of guidance mode .
In a system having at least one AGV capable of ordinarily navigating without any guidewires in the f loor 30 between t~rmin~ and of positioning itself accurately at t-~rmin~l~, an inventive object is to provide t~ninAl-positioning apparatus of a type that~ enables the same AGV
to operate also in hybrid factory installations that have some guidewires in the floor; the terminal-positioning 35 portions of the guidance apparatus are utilized for the additional purpose of following the guidewires in the f loor in order to navigate between stations .
... . _ . .

WO 92/09941 2 0 9 ~ 4 4 2 Pcr/us9l/08892 It is another object to provide AGV t~ nAl-positioning ~ LIlLu~ that enables an AGV to operate in a hybrid installation that has active guidewires (i.e., guidewires energized by conductive connections) at the floor within some of its terminals, and that has passive loops ( i . e ., conductive loops energized by magnetic induction) at the f loor witbin others of its terminals .
It is another object, in one ~-~ho~ , to utilize one or more phase-locked loops to process signals received by receiving antennas in detecting a passive conductive f loor loop, and in which the receiving circuits have ~ArAhility for initialization and for automatic gain control of the phase-locked signal level.
It is another obj ect to provide navigation and guidance apparatus for measuring, with improved accuracy, the position OI a vehicle relative to a known marker at the ~loor to ascertain the vehicle ' s position relative to a f actory ref erence system .
It is another object to utilize a generally transverse array of sensors on the vehicle to sense the marker and to process the sensed data regarding marker position in a particular way to determine the relative position of the vehicle with i uv~:d accuracy.
It is another object to determine intermittently placed guide path marker-AGV relative position by taking readings with a plurality of intermittent guide path marker sensors, including the sensor having the greatest reading, the two sensors immediately on one side of it, and the two sensors immediately on the other side of it, and correlating and interpolating the readings with a stored spatial pattern of magnetic f ield strength whereby the determination of the relative position of the vehicle with i vv~:d accuracy is realized.
It is another object to ascertain the longitudinal position of the vehicle by means of the intermittently placed guide path marker by sensing the __ _ _ WO 92/09941 ~ ~ 9 5 4 42 12 PCr/US91/08892 oc~;uLLè1~ce of maximum readings of the sen60rs as the vehicle passes over the marker.
It is another object to utilize the generally transverse array of sensor6 to CVII~:UL - e ..l_ly sense two closely positioned markers having pre~etPrm;nP~ relative and factory reference locations and, ~hereby, provide ;uLLelll_ meaciuL~ ~s for ascertaining the attitude of the array of sensors, associated bearing of the vehicle in addition to d~tP~m;nAtion of lateral and longitudinal vehicle position.
It is a chief obj ect to provide a navigation and guidance system which comprises redundant mea~uL~ I 5 of AGV position and bearing.
It is another chief obj ect to provide a navigation and guidance system which comprises guide-wire sensing apparatus, wheel ~nco(l;n~ apparatus, and update marker sensors, thereby providing re~ nrlAn~y of mea.uL L.
It is another chief object to provide a 2 0 navigation and guidance system which comprises guide-wire sensing apparatu6, wheel ~n~-o~l; n~ apparatus, angular rate sensing apparatus, and update marker sensors, thereby providing increased r~ n~n~-y of mea~uL~
It is another chief object to provide a navigation and guidance system which comprises angular rate sensing apparatus which further comprises an inertial platform which rotates to follow angular travel of the angular rate sensor whereby the maximum of f set of the angular rate sensor is reduced and gain of the output sensors are increased, thereby increasing angular rate sensor sensitivity without saturation.
It is another chief object to provide a navigation and guidance system for an AGV which comprises a one dimensional angular rate sensor.
It is another chief object to provide a navigation and guidance system for an AGV which comprises a low cost, long lived angular rate i ensor.
. .
=

~ ~ 42 WO 92/09941 ~ 4 - PCr/lJS91/08892 ;~ 13 It is a key object to provide frequent calibration of drift for a low cost, highly reliable AGV
navigation and g~ nre system angular rate sensor, which may have a high drift rate, such that an error having a standard deviation of not greater than 2 inches results in fifty feet of AGV travel.
It is another key object to use a Kalman filter to quantify values for the LL~:~UI~ drift calibration.
It is a significant object to provide guide-wire and update markers sensors suf f iciently longitudinally symmetrically ~ ~q~oFP~ such that the AGV
may operate bidirectionally as well as unidirectionally.
It is a further si~n;~ir~nt object to provide guidewire and update marker sensing which allows correction for any offsets from ,y ~Ly such that the AGV operates bidirectionally.
It is an i l.c.llt object to provide an onboard AGV traf f ic controller which comprises at least one digital processor whereby Ir-E~.gP~ received from the AGV
communication system are processed, navigation and guidance parameters are calculated, communication and sensor control switches are controlled, and rate and direction of vehicle travel and function is regulated.
It is another important object to provide a navigation and guidance system which comprises a PL~YL ~ "E~ stop, which is triggered by a digital processor interrupt and brings an AGV to a slow, controlled stop when an . ~n~:y stop requirement is detected .
It is another important object to provide navigation and guidance system which comprises a backup to the "E" stop which immediately halts progress of the AGV when an l~E~ stop malfunction is detected.
It is another important obj ect to provide a navigation and guidance system comprising an onboard digital yL O~:~S~::ur which calculates a guide path derived from current position and bearing of the AGV and the : 20 ~5442 __ ~
WO 92/09941 : PCI`/IJS91/08892 target position and bearing at the end-of-next-path 6egment received ~rom the AGV controller~.
It i6 another important object to provide a navigation and ~ ne 6y6tem which calculàte6 a guide 5 path which compri6e6 one ~PrPntlPnt and one ~ nrlQrPnAQnt variable and which i6 in~ of vehicle 6peed.
It is another ; , ~ L~ L obj ect to provide a navigation and guidance 6y6tem which calculates a non-linear guide path def ined in Cartesian coordinates .
It i6 another important object to provide a navigation and guidance 6y6tem which calculate6 a non-linear guide path defined in polar coordinate6.
It i6 another important object to provide a navigation and guidance 6y6tem compri6ing a program in 15 the onboard digital pLuce6sur which 6elect6 between and sequence6 implementation of calculated carte6ian and polar coordinate, non-linear guide path6 along which the vehicle i6 guided.
The6e and other obj ect6 and f eature6 o~ the 20 present invention will be apparent from the detailed de6cription taken with reference to 2r- ,~nying drawing6 .

WO 92/09941 ~ PCr/US91/08892 15 ~095~2 RRrFF DESCRIPTION OF 'I`T~F FIGURES
Fiqure l. Overview of an illustrative application of a preferred ~ L of the invention, including v~hi~ Ar routes and some tDrmin~l~ for pickup 5 and delivery.
Fiql1re 2. Perspective view of an automatic guided vehicle.
Fiq1~re 3. Simplified top view of the vehicle and of a passive loop of conductor in a f loor mat at a t~rmi nA 1 .
Fiqure 4A. Simplified electronic block diagram of a guidance system for a vehicle which operates in both a general purpose and a terminal-positioning mode.
Fl~lre 4B . Simplif ied block diagram providing an overview of interconnections of mnjor subsystems which operate in a t~rmin~1-positioning mode.
Fiqure 4C. Simplified electronic block diagram similar to Figure 4A, but showing only elements that are used when the apparatus is in the terminal positioning mode of operation and omitting other elements.
Fiqure 5 . Simplif ied block diagram of certain ~_I.ents of a vehicle navigation and guidance system on a vehicle for transmitting a magnetic field when operating in the tP~min~1-positioning mode.
Fia~1rç 6. Circuit diagram of an oscillator, switch, drlver, and transmitting antenna of Figure 5, which are transmitter portions of the pref erred vehicle navigation system when operating in the t~rm;n~l-positioning mode.
3 0 Fiq11re 7 . A circuit board layout showing antennas f or lateral and wire-cross positioning operations .
Fiqllre 8. Vertical sectional view of a conductor on the f loor and a receiving antenna assembly on a vehicle that is centered above it.
Fiqure 9. Another vertical sectional view of a conductor on the f loor And a receiving antenna assembly WO 92/09941 PCr~US91/08892 ~ 5~2 16 that is offset laterally above it.
Fiallre lOA. Graph of amplitude6 of signals received at magnetic receiving antennas on the vehicle as ~ function of the vehicle ' 8 lateral location relative to a current carrying wire (such as a part of a conductive loop) on the floor.
Fiallre lOB. Graph of amplitudes of signals seen in Figure lOA showing a lateral offset used to control the vehicle'6 lateral position relative to the current carrying wire.
Fi;rllre 11. Plan view of an alternative configuration of antennas and a passive loop arrangement having two turns.
Fiaure 12A. Circuit diagram of receiving antennas, preamplifiers used in common by several circuits. Also shown are rectifierg for torm;nAl-positioning operation in which only a passive loop of wire is on the f loor .
Fiaure 12B. Circuit diagram, continued from Figure 12A, of antenna output signal conditioning circuits for vehicle front-end t~orminAl-positioning operation in which only a passive loop of wire is on the f loor .
Fiqure 13. Block diagram showing an automatic guided vehicle controller (AGVC), microprocessors, and some equipment for operation in a passive wire loop mode.
Fiaure 14. Block diagram of equipment for guidewire-tracking mode of operation of the vehicle.
(See Figures 15-17 for details. ) 3 0 Fiaure 15 . Diagram of circuits including the receiving antennas, their preamplifiers, and short-circuitable attenuators ( input portion of circuit) as used when the vehicle is relying on an active guidewire for position information.
Fiaure 16A. Circuit diagram of a bandpass-filtering and signal-rectifying portion of the equipment for a guidewire-tracking mode of operation (middle WO 92/09941 ~~ PCI`/US91108892 17 209~442 portion of circuit).
Fiaure 16B. Circuit diagram, a continuation of Figure 16A, of a smoothing and ~ ator portion of the eqn~ ~ for a guidewire-tracking mode of operation.
Fiqure 17A. Circuit diagram, a continuation of Figure 16B, of a portion of an analog board that sums a command from the motion control mi.;Lv~L-,cessor with a ted error signal, and drives a motor controller.
Ficr--re 17B. Circuit diagram, a continuation of lo Figure 17A, of a portion of an analog board that controls direction of the vehicle (forward or reverse), and driYes a motor controller.
Fiallre 18. Simplified diagram of WiIL ~,L~,ssing detection circuits. (See Figure 19 for detailG. ) Fiqure 19. Circuit diagram of wire-crossing detection circuits including antennas (i.e., coils) and signal-combining circuits.
Fia~re 20. Circuit diagram of a portion of wire-crossing detection apparatus tuned to a f requency assigned for active guidewire operation of the vehicle.
Fia~lre 21. Circuit diagram, a continuation of Flyure 20, of a portion of wire-crossing detection apparatus tuned to a frequency for active guidewire operation .
Fiaure 22. Circuit diagram of a portion of wire-crossing detection apparatus tuned to a frequency assigned for passive wire loop operation in a terminal.
Fiqure 23-27. Signal waveforms at various points in the wire-crossing detection circuit of Figure 21, namely at terminals 253, 257, 267, 271, and 261, respectively .
Fiaure 28. Block diagram of an alternative clmho~ L of the invention that uses pha6e-locked oscillators in a portion of the system for processing signals from lateral-position-detecting antennas.
Fiaure 29. Block diagram of a phase-locked oscillator having automatic gain control, used in Figure , WO 92/09941 ~ 0 9 ~ 4 PCr/US91/08892 ~ 18 28 .
Fiq11re 3 0 . Plan view 6howing an alternative ~ 'ir ~ having different transmitting antenna locations on a vehicle and a passive wire loop on the 5 ground at a t~rmin~l, in which the two lobes of the pas6ive wire loop are in a side-by-side configuration.
Fia1]re 31. A circuit diagram reproducing circuits from the top line of Figure 12B and showing thereto connected circuits for calibration of an lO automatic offset adjust which ~ -ates for offsets in antenna null mea~UL~ - 5.
Fiaure 32 depicts a guided vehicle 6ystem that utilizes an update marker guidance system.
Fi11re 3 3 shows an update marker magnet in the 15 floor.
Fia11re 34 shows an array of Hall magnetic sensors on the vehicle.
Fia11re 35 is_a curve o~f analog voltage output from one of the Hall sensors as a function of distance of 20 the sensor from a floor magnet.
Fial1re 3 6 is a block diagram of some electronic equipment on the vehicle for proc~si n~ magnet sensor signals .
Fia11res 37. 37A. 37B. 37C, and 37D are, in 25 combination, a schematic diagram of the same electronic equipment .
Fia11re 38 i6 a simplified flow chart of an algorithm for prQc~C~in~ sensor data to measure the lateral position of the vehicle relative to a magnet and 30 to detect when a row of Hall sensors crosses the magnet.
Fiaure 39 i6 a simplified flow chart similar to Figure 38 including 6ensor null mea~uL. ~ and related calibration during the WAIT LOOP.
. ~

WO 92/09941 2 0 9 5 ~ 4 2 Pcr~US91/08892 Flql-res 40. 40A. and 40B comprise a simplified flow chart of an algorithm for processing sensor data to rnnrl~rr~ntly measure the lateral position of the vehicle relative to two magnets and to detect when the row of 5 Hall sensors crosses each magnet.
Fiq lre 41 is similar to Figure 34, showing an array of Hall ~ n~tic sensors on the vehicle, and ; nrl ~ ; n~ indicia exemplary of the presence of two magnets .
Fia-~re 42 is similar to Figure 32, depicting a guided vehicle system that utilizes the update marker guidance system and showing the presence of two magnets on the path ahead of the vehicle.
Fial~re 43 is a simplified top view of the 15 vehicle showing relative positions of wheels and travel measuring encoders.
Fiq~re 44 is a simplified block diagram representation of a wheel travel measuring encoder.
Fi~lre 45 is a graphical drawing of internally 20 rro~llred waveforms of a travel measuring encodQr.
Fiqure 46 is a simplified block diagram showing interconnections between wheel travel encoders and outerloop motion control proces60rs.
Fiqure 47 is a top view of the vehicle in a 25 factory frame showing factory frame to vehicle fixed f rame and inert i a 1 table relat i nn F:h i r S -Fiure 48 is a graph showing relative geometrybetween the factory frame and a waypoint frame.
Fiqure 4~ is a graph showing geometry of a 30 calculated path from a current vehicle position (and direction) to a next waypoint and direction of travel along the :~hc~r; ~ 21 of the graph.
Fiqllre 50 is a graph showing ge ~Ly of a travel segment from a present vehicle position to a 35 waypoint involving circular travel.
Fiq~lre 51 is a graph showing g~ l_Ly of circular travel of a vehicle in a vehicle frame to a .

-WO 92/09941 2 û ~ ~ 4 4 2 PCr/US91/08892 ~

waypoint in a waypoint frame.
Fial-re 52 i6 a graph showing a relationship hetween length of travel and a speed setpoint, from which a calculation is made to change the speed of a vehicle as a function of a length of travel.
Fiaure 53 is a simplified graph showing the geometrical relat;r~n~hipc between the path between two markers and the factory frame.
Fial1re 5~ is a graph showing the path geometry for a manually guided path from which measurements are made relative to vehicle insertion into a factory frame.
Fiaure 55 is a simplified block diagram o~ an inventory management system showing interconnections between a vehicle controller (AGVC computer) and a ~-nr ~ system controller.
Fiaure 56 is a block diagram of a vehicle navigation and guidance system showing relationships between outer and innerloop (motion control processor) control elements.
Fi~lre 57 is a simplified perspective of an inertial platform for the vehicle.
Fiaure 58 is a block diagram of a model of the StAhi 1 i ~Ation and control loop of the inertial platform.
Fia--re 59 is a circuit diagram for the stabilization and control loop of the inertial platform.
Fia~-re 60 is an assembly drawing of the inertial platform with parts cut away for clarity of presentation .
Fia-lre 61 is a simplif ied model of the heating system of the angular rate sensing element of the inertial platform.
Fiaure 62 is a plot showing curves from sensors used in the feedback loop in the angular rate sensing element of the inertial platform.
Flaure 63 is a block diagram of the inertial platform which is part of the stAhi l i 7~tion and control loop .

WO 92/09941 ~ 2 0 ~ ~ ~ 4 2 PCr/~S91/08892 ~ Fi~lre 64 is a diagram of a "straight line"
guidepath showing relationship of a waypoint frame to the f actory f rame, Fiql~re 65 is a diagram of an "arc" guidepath 5 showing a plot of a turn in a waypoint frame.
Fiqllre 66 is a diagram showing a family of three guidepaths each of which results from a different position of the vehicle with regard to a waypoint frame.
Fiqllre 67 is a diagram showing a family of 10 three guidepaths where a turn of the vehicle is executed from different relative positions in a waypoint frame.
Fiql~re 68 is a diagram of a complex guidepath formed by syccessively calculated guidepaths comprising in seriatim a "straight line" and then an arcing or ~5 curved guidepath.
Fi~llre 69 is a schematic showing a magnet and the relative position of a sensed field that ~ULL~UIIdS
the magnet and points comprising time delays whereat the vehicle controls recognize the sensing of the magnetic 20 field.
Fiq~lre 70 is a top view of a vehicle having LL.-v~l~ed a ground marker from which a position mea~.uL I (. has been made showing some sources of errors comprising the offset of the ground markers from the Z5 centerline of the vehicle and delays due to motion of the vehicle after the marker is sensed.
Fiqllre 71 is a simplified block diagram of a wireless i cation system showing source of control from an automated guided vehicle controller and two-way 30 communications between the controller and a related base station and at least one vehicle electronics.
Fiqllre 72 is a block diagram of base station communication electronics and a related radio.
Fiaure 73 is a block diagram of vehicle 35 communications electronics showing connecting relationships among a ~ i cations processor, an SDLC
chip, a radio data decoder, and a two-way radio.

WO92/09941 - 2 0 ~i 5 4 4 2 Pcr/US9l/08892 ~
Fiaure 74 is a 6chematic of the circuits for the radio data decoder, and communications control and data lines comprising request to send, clear to send and permission to transmit, transmit data to audio conversion, transmit clock control, and power regulation circuits .
- Fiaure 75 is a detailed schematic of the automated guided vehicle controller ~ tions electronics which receive input from the radio data decoder, said electronics comprising an SDLC chip, a central processing unit, a clock generator, and floor controller interfacing circuit$.
Fiallre 76 is a timing diagram showing representative wavef orms involved in practicing the radio data decoder.
Fiqllre 77 is a simplified block diagram showing the plurality of mi~:rvpIoc~ssv~S used in guidance and control of the vehicle, interconnecting bus lines between the processors, and some of the input devices which 2 0 connect to the EJL vCeS5U.~
Fiaures 78. 79. 80 81. 82. 83. 84 and 85 provide a schematic of the ~ -nr-ntS and circuits of a communications board, comprising two central processing units, contained in each vehicle.
Fiallre 86 is a map showing relative orientation of the schematic of circuits seen in Figure 92 to the schematic of circuits seen in Figure 93.
Fiaures 87. 88. 89. 90. 91. 92. 93. 94 and 95 provide a schematic of a digital I/O board, comprising five central processing units, contained in each vehicle.
Fiallre 96 is a graph showing a proposed path for a vehicle in a waypoint frame wherein measurements are made to determine p ~th selection.
_ _ _ WO 92/0994l 2 0 9 5 4 4 2 Pcr/Us9l/0~892 DE~ATT T'n DEsrRTpTIoN OF 'rTTT.' TT.T.TT~TRA'rEn T~MT~DTMT`NTS
In this description, two sets of terms are used to reference angular direction of travel of an automatic guided vehicle (AGV), port and starboard and left and right. Port and starboard are directional references of left and right, respectively, to the true vehicle front which may be identified by the presence of a light, rl~ L of a grill, indicia, or other At ~-Da~:~ry at the front of the AGV. Right and left are references with regard to the direction of travel of the AGV. As an example, because the vehicle operatively travels both forward and backward, port is left when the vehicle is traveling in the direction of the true vehicle front and right when the vehicle is traveling in the reverse direction. Reference is now made to the: _~;r-nts illustrated in f igures 1-95 wherein like numerals are used to designate like parts throughout.
ovPrview of An hUt( 'ic G~ ed Vehicle ~AGV) Control ~çm , . =
The AGV control system comprises an automatic guided vehicle controller (AGVC), at least one of a plurality of types of guide path marking systems, at least one AGV comprising navigation and guidance systems capable of operating over the plurality of types of guide path marking systems, and a two-way communication system between each AGV and the AGVC. The plurality of types of guide path marking systems and AGV navigation and guidance systems comprise guidewire marking and navigation and guidance, such that new AGV's and AGVC's 3 0 are downward compatible wLth current guidewire installations .
In Figure 1 the interior of a warehouse h~ l;n~, in which automated guided vehicles, generally designated 2A, travel about on routes such as routes 3 and 5 among a number of tD~m;nAl~ such as tDrm;nAl ~ 9 and 11, is schematically shown. This is an example of a hybrid facility. The routes 3 have guidewires in the .

_ . . _ .

WO92/09941 S- 2 0 Q 5 4 ~ 2 ~ PCr/usg1/08g92 ~

f 104r to def ine the routes and guide and _ ; ~q~te with AGV'6 travelling thereon. The routes 5 are traver6ed by self-contained navigation and g~ lA~ e and wireless ; cating AGV ' 8 which f ollow paths marked by update markers 6 located at irregular intervals as much as 50 feet apart along the routes 5. The same vehicles are used on both types of routes. Routes 3 shown a6 6ingle wires in Figure 1 r.:~r~a6~ guidewire loops as is well known in the art . Each guidewire receives power f rom AGVC 13. Update markers 6, constituting, in combination, guide path 5, L~:yLesel,L devices from which accurate positioning may be derived and which may be magnets as described in detail later.
Referring to Figure 55, AGVC 13 comprises an AGVC computer 13A and at least one f loor controller 13B, and may further comprise at least one ~, oyL hle logi~
controller (PLC 13C). AGVC 13 software is currently commercially available in AGV 2A guidewire systems sold and distributed by Eaton-Kenway, 515 East 100 South, Salt Lake City, Utah 84102. As seen in Figure 55, the AGVC
comprises a ; cations link whereby a management computer 13D is connected and through which tasks are assigned to AGVC 13. Thus, AGVC 13 may be part of a larger inventory management system 1000 which is controlled by rqn~3 L computer 13D. In addition to controlling AGVC 13, r-nq,~ ~ computer 13D processes orders, maintains an inventory, produces reports, and manages C~JIlv~ r tracking and operation of vertical stacker controllers 1002 whereby material is moved to stacks and retrieved from within a storage facility 1004.
Upon receipt of load ~ ~ task from manay~ L computer 13D, AGVC computer 13A selects an AGV
2A and Erh~ e an optimum path for the selected AGV 2A.
Based upon the path scheduled and the current position of each AGV 2A, AGVC computer 13A proYides path segment by path 6egment control of - ~ . l. of each AGV 2A under its control through two-way _ ; c~tions between the AGVC

;` ~
WO 92/09941 ~ ~

and each of the AGV 18 2A . The leng~h of each path segment range6 from a fraction of the length of an AGV 2A
to a length greater than an AGV 2A length, which can be a plurality of AGV lengths.
AGVC 13 provides signals ~o the vehicles via guidewires Pmhe~ Pd in the f loor when they are operating on the routes 3. As described in U.S. Patent 4,791,570, AGVC 13 can communicate with a plurality of communication circuits each connected to a guidewire. As seen in Flgure 1, it also sends the same dal:a through wireless mtenna 15 f or vehicles not on a guidewire path or otherwise unable to receive communications from AGVC 13.
In order to keep transmission and reception from AGVC 13 and each AGV 2A, mutua~ ly exclusive in the currently preferred ~ L, a communications protocol has been adopted f or both the guidewire and wireless modes of communication. The protocol gives priority to transmission from AGVC 13 such that no data is transmitted from any AGV 2A when the AGVC 13 is transmitting. All data transmitted by AGVC 13 is transmitted globally, that is, it is transmitted by each ~nd every communication circuit in the system. To avoid data collision, each AGV 2A only attempts to transmit data when it has been polled by AGVC 13.
In addition, AGVC 13 monitors obstacle (such as fire control, exit doors, etc. ) and other discreet devices, AGV 2A battery status, sizing mea~uL, Ls in sizing stations and controls site specific devices such as lights. All monitoring and controlling is performed 3 0 over both hard wired and wireless communications, as available .
~hP AGvc ~- i catinn~ SYstem AGVC 13 comprises multiple communications modes. As seen in Figure 1, an AGV 2A can travel over a guidewire route 3 or a ground marker route 5. When AGV
2A travels over guidewire route 3, communicating -Ca~PC
can be used over the guidewire or via wireless WO92/09941 ~ 2 ~ ~ ~ 4 4 2 PCI/US91/08892~
i cation6. When AGV 2A traverses a path 5 of ground markers 6, a wireless ir~tions system 111~ (see Figure 71) is used. In a hybrid facility comprising both guidewire and ground marked paths where at least one AGV 2A may be on each route 3, 5 at any time, AGVC 13 must i, Ate over both modes col~-.u~ Lly. The circuits ~nd methods f or i cAting over a guidewire are the same as those described in U.S. Patents 4,491,570 and 4,902,948, which are the ~-up~:-Ly of the assignee of this 10 invention and which are made a part hereof by reference.
As seen in Figure 71, the wireless communications system 1110 comprises a non-vehicle portion 1100 and a vehicle portion 806. The non-vehicle portion 1100 comprises AGVC 13 which includes an AGVC
computer 13A interconnected to base station 802 by either an RS422 or an RS232 communicating link 828. Base station 802 is electrically connected to a radio 804 which sends and receives through an antenna 15 whereby wireless communications are sent ~o and received from the plurality of vehicles 2A in the facility. In the currently preferred ~Tnhor~ir L, radio 804, commonly used by both base station 802 and each AGV 2A, is a model I~S-900, available from TE}~K Inc., 224 N.W. Platte Valley Drive, }~ansas City, No. 64150, although other radios can be used within the scope of the invention.
A block diagram of base station 826 is seen in Figure 72. Base station 826 comprises circuits 808 which selectively convert RS422 and RS232 signals to levels E~rUces5~:ble by base station 826 logic circuits. Circuits 808 communicate with a central processing unit 810 via output line 814A and input line 814B. Although other central processing units can be uqed within the scope of the invention, central processing unit 810 is a DS5000 (from Dallas Semiconductor) in the currently preferred e~-~o~ . Central processing unit 810 communicates with an SDLC chip 812 which operates in the same manner as SDLC chips ln guid_~re commur ic.-tions. Output from ~O 92/09941 ~ 4 ~ 2 Pcr/us9l/08892 SDLC chip comprises request to sent (RTS 840), and transmitted data (TxDATA 850) to a radio data decoder 820. Inputs to the SDLC chip from the radio data decoder 820 compri6e a transmit clock tTxCLK 852), a clear to send (CTS 842) signal, and received data (RxDATA 874).
Circuits and operation of radio data decoder 820 is A; cc~c~ed in detail hereafter. Radio data decoder 820 is cnnn~cted to radio 804 by an audio transmit line (TxAUDIO 866), a permit to transmit (PTT 868), and an audio data receive line (RxAUDIO 870). Thus data i5 received from AGVC 13 in RS232 or RS422 format, transferred to, buffered in memory, and resent from central processing unit 810 to the SDLC chip 812. From SDLC 812, data is sent to decoding circuits, wherein the data is translated for efficient transmission, and therefrom sent to radio 804 for transmission through antenna 15.
Base station 802 received data is detected at antenna 15 and relayed to radio 804 wherefrom, the data in audio format, is sent to radio data decoder 820 wherein the audio RxAUDIO 870 signals are transformed to RxDATA 874 signals which can be processed by SDLC chip 812. Once processed by SDLC chip 812, data is sent through bus 816 for storage in memory and further transmission to AGVC computer 13A after conversion to the selected RS422 or RS232 format.
The message format sent via wireless transmission is the same as the ~ormat described in the earlier referenced U. S. Patents ' 570 and ' 948 . The message format being:
<BOM><AGV 2A address><message><CRC><EOM>
The message within the message f ormat is either a command or status and may be of any length. BOM and EOM are the same beginning and end of message codes used in guidewire communications. The CRC check code is also calculated in the same manner as the CRC in guidewire communications.
h ~ ~ ~ por~ on ~06 ~' wlreless WO 92/0994~ ~Q ~ PCr/US91/08892 ; c-atiOns system 1110 is seen in Figure ~3 . Intra-vehicle c~nnnpc~t1 ~7nc are not shown but are identical to those described in previously referenced U. S . Patents ' 570 and ' 948 . Inte~c~ e-;-ions and operation of central proGP~ein~ unit 810', SDLC chip 8;L2, and radio data decoder 820 are the same as the same central procP~ein~
unit 810 ', SDLC chip 812, and radio data decoder 820 used in the non-vehicle portion 1100 of i c ntion system 1110 . Though not nPc .~c5~ r y within the scope of the invention, the same radio 804 is also used.
A digital dec~orlin~ circuit 1120 portion of radio data decoder 820 is seen in the circuit schematic in Figure 74. A 9600 baud digital data stream is sent to radio 804 wherefrom the signal is modulated and sent over a carrier wave to another receiving radio 804. Using digital dP~o~lin~ circuit 1120, the 9600 baud data stream requires a ba6e band of only one-half the 9600 baud digital data #tream freguency to ~end a signal which, as received and provided by a receiving radio 804, produces a diseriminator waveform seen as discriminator output 1136 in Figure 76. Digital deeo~lin~ eireuit 1120 reeeives and reeonstruets the 9600 baud signal whieh is transmitted effeetively at 4800 eyeles per seeond. Even at 4800 eyeles per seeond, reeeived signal amplitude is substantially lower than other radio signals whieh are transmitted at frequencies lower than the 3000 cycle per second base band cutoff of radio 804. Digital decoding circuit 1120 is of primary importance in the digital data rec.,~.D~ uuLion beeause a frequeney of 9600 eycles per second is too far beyond the 3000 cycle per second base band cutoff of radio 804 to be reliably detected. Even 50, the amplitude of the 4800 eyele per second frequency signal requires special processing to reliably r eC ullL ~L uct the original digital data stream .
Digital ~lPco~in~ circuit 1120 is similar to the circuit tlie--loePd in U.S. Patent 4,613,973 which is the property of the assignee of this invention. Input to , WO 92~09941 2 0 9 ~ 4 4 2 PCI/US91/08892 circuit 1120 is RxAUDI0 870 which iB received from radio 804. The digital cl~ro~l;n~ circuit 820 produces a digital signal which i8 _ i~Rted to SDLC chip 812.
Digital ~loco~;n~ circuit 1120 as presently 5 preferred, comprises seriatim a differential amplifier 828, a comparator circuit comprising a positive ~ tor 830A and a negative comparator 830B, one digital level translator 832A, 832B for each comparator 830A, 830B, and a latch circuit comprising a flip-flop 1122 formed of two inverting AND gates 834A and 834B.
Differential amplifier 828 produces distinct voltage spikes CULL _L- ,-l;n~ to voltage transitions of the waveform received from radio 804 across a zero voltage.
A positive voltage spike is produced whenever the waveform passes from negative to positive, and a negative voltage spike is produced whenever the wavef orm pas6es from positive to negative.
The comparator circuit produces a voltage at each comparator 830A and 830B, a first voltage which is interrupted whenever the output of differential amplifier 828 exceeds a certain predetermined value and a second voltage which is interrupted whenever the output of the differential amplifier 828 is less than a certain pre-det~rm;n~d value.
Digital level translators 832A, 832B are MC1489 chips (from Motorola Semiconductor) which are more generally used in RS232 positive/negative voltage levels to digital voltage levels conversion. In this case, the input levels to digital level translators 832A, 832B are 3 0 + 12 volts; output is compatible with standard TTL
voltages .
The latch circuit receives the outputs of digital level translators 832A and 832B as set and reset inputs, respectively, to the flip-flop 1122 thereby producing as an output a digital data stream.
Referring to Figure 74, the output of radio 804 is provided with a load resistor R73C, chosen to balance -WO 92~09941 ` t~`~ O 9 5 4 ~ 2 PCr/US91/08892 ~

the capacitively coupled output of discriminator circuitry contained in radio 804. A representative output 1136, seen in Figure 76, from radio 804 comprises a digital data stream 1134 which has been distorted by 5 modulation and demodulation of a carrier wave. The output 1136 is pa66ed to differential amplifier 828 which comprises a differentiating input through capacitor C207C. The differential amplifier used in the currently preferred embodiment is TL072 from Texas Instrument. A
10 list of components used in the currently preferred embodiment is found in a table below. It should be understood that the components used in the list are for the currently pref erred ~mho~ t and other components can be used within the scope of the invention. The other resistors R74C and R80C and capacltor C205C function in a known fashion. The differential amplifier 828 operates as a differentiation device by resistor R74C and capacitor C207C. The feedback resistor R80C and capacitor C205C are provided for the purpose of limiting 20 input bandwidth to suppress high frequency noise.
Given the data stream 1134, the output of differential amplifier 828 comprises a waveform 1138.
See Figure 76. The maximum amplitude of wave~orm 1138 is adjusted to a suitable value for example, in excess o~
25 6 . 5 volts positive and negative, by either adjusting the amplitude of waveform 1136 at the output of radio 8~4, or by choosing appropriate values for the resistors and capacitors used with amplifier 828.
The output 1138 of the differential amplifier 30 828 is communicated to the comparator circuit comprising positive voltage comparator 830A and negative voltage comparator 830B. Comparators 830A and 830B are provided with suitable comparison voltages through voltage dividing resistors R81C, R82C, and R79C and pull-up 35 resistors R83C and R78C. A satisfactory integrated circuit is an I~339 available from National _ _ WO 92/09941 2 0 9 ~ ~ ~ 2 PCI /US91/08892 Semiconductor. In the preferred l~rho~ nt, the comparison voltages are provided by a source of positive potential and a 60urce of negative potential connected by a voltage divider formed by R81C, R82C, and R79C. For 5 the + 12 voltage level used in the pref erred c~nho~ i r -nt, re6istor value6 are tho6e li6ted the table below.
Digital level translators 832A and 832B are inserted between comparators 830A and 830B, respectively, and produce logic level outputs f or the inputs to f lip-flop 1122.
The function of positive comparator 830A is tointerrupt a current at the output of positive comparator 830A whenever and as long as the amplitude of waveform 1138 exceeds the positive comparison level voltage. A
repres~:l.Lcltivt: output 1140 from positive comparator 830A
is seen in Figure 76.
The function of negative comparator 830B is to interrupt a current at the output of negative comparator 830B whenever and as long as the amplitude of waveform 1138 is less than the negative comparison voltage. A
representative output 1142 from negative comparator 830B
is 6een in Figure 76.
The two output6 of comparator circuit6 830A, 830B are respectively applied to the inputs of flip-flop 1122 through digital level translators 832A and 832B, respectively . In the currently pref erred ~ho~ nt, a set-reset flip-flop 1122 comprises two negative logic AND
gates (NAND gates 834A and 834B) connected as shown in Figure 74. A 6uitable negative logic AND gate i6 SN74279 available from Texas In~,LL, - Ls. A low logic level at the input of NAND gate 834A will set flip-flop 1122 output to a logic level "high". Flip-flop 1122 output is reset to " low" by a " low" logic level at the input of NAND gate 834B. Flip-flop 1122 outpu~ 1148 voltage level 35 (Figure 76) represents a stream of digital data corr~pr~n~lin~ to the digital input 1134.
The values of circuit _ ^nt6 are not . _ =

WO92~09941 32 PCr/US91/08892 critical to the operation of receiving circuit 1120. Of course, the combinations of resistors and capacitors are chosen such that the response time (or "time con6tant") of the differentiating circuit is compatible with the 5 input rL-:yuu,l~y from the radio 804. Variation6 and modifications may be made without departing from the present invention. In operation, receiving circuit 1120 provides a rapid "off" to "on" time of less than five mi 11; ~::Pcnn~
A dPt~ P~ schematic of the circuits comprising intercnnnPct; t~n~: between central processing unit 810 and SDLC chip 812 is found in Figure 75. Connections between each of the ~ s are standard and known in the art.
Therein, RxDATA 874 is received through switch E15 to the 15 RxD input of SDLC chip 812. Clock generation is provided by oscillator 865 and clock divider chip 864. Light emitting diodes DSlAC provide visual status of operation of SDLC chip 864. All "E" references specify computer controlled switches or jumpers.
Of particular interest is the interface to a guidewire floor controller which comprises the interfacing circuits 1150 enclosed by dashed lines in Figure 75. Interfacing circuits 1150 comprising guidewire floor controller drivers of transmit driver 1034, receive amplifier 1032, transmit clock 1030, and an output amplifier for a sixty-four times clock 1028.
C -nts used in currently preferred ~ `--'; - L of the wireless ~ ; cation system as seen in Figures 74 and 75 are found in the following list:
3 0 ~k~ Name V~ 1 llP or Tvme R2C Resistor 2 . 2K Ohms R3C Resistor 100 "
R4C Resistor 2 . 2K "
R5C Resistor 2 . 2K "
R6C Resistor 2 . 2K "
R8C Resistor 100 "
R9C Resistor 2 . 2K "

WO 92/09941 - _ 2 0 9 ~ 4 4 2 Pcr/US9l/08892 RlOC Resistor 4 . 7K
R14C Re6istor 2 . 2K
R15C Resistor 180 R16C Resistor 180 5 R17C Resistor 180 "
R18C Resistor 180 R73C Resistor 30K
R74C Resistor lOK
R77C Resistor lOR
10 R78C Resistor lOK
R79C Resistor 3 . 3K
R80C Resistor lOOOK "
R81C Resistor 3 . 3K
R82C Resistor 5 . lK
15 R84C Resistor lOK
R83C Resistor 100 R9OC Resistor 4.3K
R9lC Resistor lOK
R92C Resistor 100 "
20 R93C Resistor 2.4K
R2 01 C Res i stor 2 0 0 R202C Resistor 1. 33K
C205C Capacitor 10 pf C207C Capacitor . 01 l~f 25 C207AC Capacitor 10 ,~f C27C Capacitor 33 pf C28C Capacitor 33 pf C6C Capacitor 100 ,uf C31C Capacitor O . 47 ~Lf C32C Capacitor 0.47 ~Lf C33C Capacitor O . 47 ,~Lf 834AC Nand 74LS279 834BC Nand 74LS279 . ~ _ WO92/~9941 ~ ~ ~ 2~ ~ 0 4 ~ a ~ PCI/US91/0~892~
836AC IDv.Amp. Std 8 3 6BC Inv . Amp . Std 838C Nand 74L5132 838AC Nand 74LS00 5 848C Inv.Amp. Std 860C Pow.Reg. LM317 864C Clock Gen. 74HC4040 1034C Out.Amp. 3487 1028C Out.Amp. 3487 1030C Input Amp. 3486 1032C Input Amp. 3486 CRlC Diode lN914 UlC Nor 74LS132 U6C Inv.Amp. 74LS04 U7C Inv.Amp. 74L504 U14C Nor 74L502 U15C Amplifier 7407 U33C Inv.Amp. 1489 U82C Diff.Amp. TL072 U83C Diff.Amp. LM339 U90C And Gate Std WO 92/09941 2 ~ 2 PCI/US91/08892 G~ Path Markinq AGVC 13 controls automatic guide vehicles over a plurality of guide paths. As seen in Figure 1, the guide paths may be a substantially continuous guidewire or series of guidewires activated by a central source such as AGVC 13, a sequence of intermittently placed update markers requiring an AGV 2A to traverse therebetween by self contained guidance, or a passive guidewire not connected to a power source but receiving emitted power induced from the AGV 2A, itself. These paths are depicted by guidewires 3 and update markers 6 in Figure 1. Passive wire loops in a mat 51 as seen schematically in Figure 3. The passive wire loop in a mat 51 provides u~u,uu- ~u~ity for guidewire guidance where there is no power connection to AGVC 13. Such u~LLullities are found in tc~rmin~l positioning and providing temporary paths between otherwise marked guide paths .
Thf~ tomatiC Guided Vehicle (AGV 2A) One automatic guided vehicle, AGV 2A, is depicted isometrically in Figure 2 and schematically, showing p~ ~I t of wheels and cactors in Figure 3 . It has drive wheels 8, 10 on its port and starboard sides respectively, which are powered individually by motors.
Casters 12, 14, 16 and 18 support the vehicle at its port front, port rear, starboard front and starboard rear corners respectively. As earlier described, the terms port, starboard, front and rear refer to physical absolutes of the vehicle. The terms left and right are 3 0 relative to the direction of travel; the vehicle operates .,y ~Lically in either direction. The front 2F of vehicle 2A as seen in Figure 2 comprises and is identified by two laterally disposed grills 2G and a control panel 2P which comprises a light 2L. The rear 2R
of vehicle 2A is the other end. Port and starboard are referenced to the front 2F of vehicle 2A. These terms are generally used herein.

_ _ _ _ _ _ _ _ . . . . . . . .

:
WO 92/09941 2 0 9 ~ 4 2 PCI/US9l/08892 ~

~rouch-sensitive bumpers 20, 22 are located at the front and rear of the vehicle, respectively, to detect obstacle& in the path and to activate switches to stop the vehicle.
In addition to the mechanical parts mentioned nbove, each AGV 2A further comprises sensors, a two-way communication sy6tem, a navigation and guidance system, and a vehicle traffic control system, each of which resides below the top surface 28 of AGV 2A wherein vehicle 2A comprises a well 26 used for navigation and guidance apparatus, leaving the top 6urface free for loads or other uses, as seen in Figure 2.
ThP AGV 2A CommunicationS Sy6tem The AGV 2A communications system comprises both guidewire and wireless communications capability.
Guidewire communications are the same as ~ closPd in U. S . Patents 4, 491, 570 and 4, 902, 948 which are the property of the assignee of this invention and which are made part hereof by reference. A block diagram of the wireless c ; r;~tions sy8tem is seen in Figure 73 . As seen therein . ; cations board 824 comprises wireless communications --ts and circuits which are similar to the wireless communications c ~-lts and circuits seen in the block diagram of base station 826 in Figure 72. However, a central processing unit 8742 is used in communications board 824 while a DS5000 central processing unit is used in base station 826. Even so, wireless ~ ~; cations functions of ~ j cations board 824 and base station 826 are the ~ame. The major difference is the higher volume message h~ntll ;n~ and buffering required of base station 826.
As seen in Figure 73, a radio 804 is located in each wireless ~ ; cating vehicle 2A and receives signals via an antenna 15. Communication lines RxAUDIo 870, PTT 868, and TxAUDIO 866 transmit received audio digital data streams, permission to transmit, and digital data streams to be transmitted, respectively, in the WO 92/09941 2 0 9 ~ 4 4 2 PCr/US91/08892 directions shown, between radio 804 and radio data decoder 820.
Radio data decoder 820 operates as earlier described. P.180 as earlier described, lines RxDA~A 874, RTS 840, CTS 842, and TxDATA ioate received data, rec~uest to send, clear to send, and data to be transmitted, respectively, between radio data decoder 820 and SDLC 812, over lines 818 in the directions shown.
SDLC operate6 a6 i5 well known in the art. A bus 816 provides ~ tion between SDLC 812 and CPU 810 ' . A
clrcuit diagram which include6 the circuits related to the vehicle 2A is provided in Figures 87-95 and hereafter described as part of the vehicle 2A ~i~:Lu~uL~cessor system .
The AGV 2A Sensors Each AGV 2A comprises a plurality of sensors and sensors types providing mea ,UL - - ~ capacity f or a plurality of guide path marking systems and redllnrlAncy of mea:,ur~ ~ whereby the effects of systematic sensor errors are dyn~micllly removed from the estimates of AGV
2A position and direction of travel. A navigation and guidance sy6tem provide6 a plurality of operating modes for guiding the AGV 2A over a number of different guide paths. As seen in Figure 3, sensor~ of the currently preferred c ` ;'i- ~ comprise antennae 47 for measuring a magnetic field emitted by guidewire 3 or a mat 51, Hall sensors 2 4 f or measuring each traversed update marker 6, which, in the currently preferred Pmho~ , comprises a magnet, as described in detail hereafter, an angular rate 6ensor system 500 for dynamically measuring vehicle direction, and an encoder 58 for each fifth wheel 57 and sixth wheel 59 for measuring travel at the port and aL-l sites of AGV 2A, respectively.
A simplif ied top view of AGV 2A is shown . onc~ually in Figure 3 . An update marker 6 is shown on the floor on the left side of the figure. This is a guidance system of the type represented by the routes 5 ~ . ~ . , , W092~09941 ~ ~ a~ ~442 ' PCrJUS91/08892 of Figure 1. In Figure 3, on the ground at the terminal 11 is a mat 51, which has a loop of wire 54 in the shape of a skewed figure eight ~ 1 in it. A left-hand portion or lobe of the loop is designated 53 and a right-5 hand portion or lobe i8 designated 55. An antenna system47 is near the front of the vehicle; it is centered on a longitudinal centerline of the vehicle and extends transversely. A similar antenna system 47A is at the rear .
Figure 3 also shows an array of Hall sensors 24 that are employed in the navigation and guidance system of the vehicle, as well as other navigation and ~~ nre subsystems and ~ Ls including a ~yLusco~e 63, a navigation computer 67, a motion control processor (computer) 61 and fifth and sixth wheels 57, 59 for measuring the travel of the port Pnd starboard sides respectively of the vehicle. In combination, these sensors provide rptll~n~lAnry of mea~lL~ ~ whereby errors due to causes comprising drift, m;~:rs~l ;hration, wear, t~ uLe: change, and variations in vehicle response to load and use are dynAmir~lly corrected. A Kalman filter 65 is used to evaluate such errors in each sensor and provide adjusting corrections to the navigation and guidance system as described hereafter.
The position-sensing po~tion of the vehicle ;nrlll~Pc a magnetic-field transmi~ter on the vehicle, the passive loop of wire 54 on the floor, and signal-receiving equipment on the vehicle. During operation of the system as a whole the vehicles 2A drive about on the various segments of the routes 3, S as shown in Figure 1 to pick up and deliver loads. ThQ vehicles are propelled f orward and steered by rotation of the drive wheels 8 and 10. ~he direction and speed of each wheel is controlled by its respective portion of a control system as described below.

WO 92/09941 ` 2 ~ ~ 5 4 4 2 PCI'/US91/08892 ~h-~ ~rv 2A Naviqation and Guidance Svstem U~date Marker Gu~fl~n~e Svstem Figure 32 is a stylized top view of the guided vehicle 2A driving in the direction of the arrow 4' toward a magnet 6 that i6 mounted in the f loor . AS
related earlier, vehicle 2A has drive wheels 57-, 59 on the left and right sides respectively, which are powered individually by motors that are not shown in Figure 32.
Casters 12, 14, 16 and 18 support the vehicle at its left-front, left-rear, right-front and right-rear corners respectively. The terms front and back are used here for convenience of description; the vehicle operates y L lcally in either direction .
T.JU~ cnsitive feelers or bumpers 20, 22 are located at the f ront and back of the vehicle respectively to detect obstacles in the path and to activate switches to stop the vehicle. A transversely arranged linear array of magnetic sensors 24 is mounted on the vehicle as shown in Figure 3 2 .
2 0 g~date Marker SYstem - The f loor maqnet In Figure 33 a floor marker 6 is shown in place in a hole 32 in the floor. In this ~ nt, floor marker 6 comprises a cylindrical magnet, placed with its axis vertical, and has its south-polarized face 34 facing upward ana its north-polarized face 36 at the bottom of the hole. Since only magnets are used in the currently pref erred ~ , the term marker 6 and magnet 6 wil be used interchangeably. However, this interrh~n~DAhle use is only for the purpose of simplicity and clarity of presentation. In the general case, it should be understood that more than one kind of f loor marker can be used in the invention. The diameter of the magnet 6 in this ~nhQdi t is 7t8 inch and its axial height is 1 inch .
_ _ _ -W092/09941 ~ 5 4 ~ 2 PCI/US91/08892 MA- n- tiC-Field Sensors The array 24 of ~-~n~ti~-field sensors is 6hown in plan view in Figure 34. In this . ;r ~ it comprises twenty-four Hall-effect sensors spaced for 5 example 0 . 8 inch apart in a straight line perp~nA i c~ r to the longitudinal centerline 559 of vehicle 2A and laterally centered on the centerline 559 of vehicle 2A.
The first sensor is labeled 437; the twelfth sensor is 448; the thirteenth Bensor i5 449 and the twenty-fourth 10 sensor is 460.
The sensors 24 are commercially available devices whose analog output voltage varies as a function of the magnetic field it detects. Each sensor has a null voltage which is its output when no magnetic f ield is 15 present. When a magnetic f ield is present the voltage consistently increases or decreases relative to the center of flux of a magnet and to the null voltage ~r~n~;n l upon whether the magnet crosses a south or north pole . In the described : - 1 i r ~ of the invention 20 the sensors always detect a south pole field 34 so their output voltage always increases as a result of being near magnet.
A repre6entative graph 464 of the analog output voltage versus distance of a sensor from the center of a 25 magnet 6 is shown in Figure 35. Voltage output from the Hall sensor (such as sensor 445 for example) is shown on the ordinate 462 in volts. The distance 145 from the center 557 of the magnet to the sensor is shown on the l~hsCi~ :a 461 in inches. For the mea~u~ L shown the 30 graph has a de:yI~sfied zero and the output voltage in the zlbsence of any magnetic f ield is the null voltage 66 of about 6 . 44 volts .
In this mea .u~ ~ when the sensor 445 is directly over the center 557 of the magnet the analog 35 output voltage is approximately 7.1 volts. When the sensor 445 is approximately one inch away from the center 557 o: ~ e ma~ne~ 6 the analos ou~put vo~age 464 WO 92/09941 2 0 9 ~ 4 4 2 Pcr/usgl/08892 ùuced by the sensor is approximalcely 6 . 65 Volts.
Thus, two magnets which are more than four inches apart, but sufficiently close to be simultaneously sensed, produce detect~hl 1~ signals which are essentially ~ ,L .
~ircuitS for Procegs;n~ Sen60r Siqnals Signals from the twenty-four Hall sensors of array 24 are input at t~rmin~l~ 468, 469 to a pair of ganged multiplexers 470, 471, as shown in Figure 36. The multiplexers 470, 471 receive analog signals continuously from the twenty-four sensors 437-460, and select one at a time sequentially for output at line 472. The two output signals from the multiplexers are connected to a signal-conditioning circuit 474 whose functions are explained in more aetail below. Its output at line 476 ls connected to an analog-to-digital converter (A/D) 478 whose output comprises eight digital lines 480 tllat conduct digital signals to a mi~:Locv.,~Luller 482.
Output data from the mi- Lucu-.LLuller 482 are in serial form differential output at a line 484, which cc,l~du~ , the data through a i cation chlp 485 and differential output Iines 481, therefrom, to a communication board, not shown. A control bus 486 enables the mi~:LuconLLuller 482 to control multiplexers 470, 471 and the A/D converter 478 as described more fully below.
~i rCuit Details More details of the electronic circuits on the vehicle are shown in Figures 37 and 37A-D. In combination, Figures 37A-D comprise a single circuit layout, ' ~d in clockwise rotatlon and dlvided as seen in Figure 37 . Int~ i ons among Flgures 37A-D
comprise twenty-four lines between Figures 37A and 37B, six lines between Figures 37B and 37C, and four lines between 37C and 37D. The line5 between 37A and 37B
comprise twenty four sensor inputs 468, 469.
Intt:L- u....e,_-iOns between 37B and 37C comprise five lines, .

~ ~ =
~= ~'`~

WO 92/09941 ` ' ~ PCr/US91708892 2~9544~ 42 generally designated 514 , and line 16 . Lines 484, 484 ', 514 ' and 631 ' connect _ n"~-nts of Figures 37C and 37D .
The twenty-four sensor inputs 468, 469 are cnnn~ct~d to two sequentially addressed multiplexers which may be Model AD7506 multiplexers. Outputs 472, 473 are each connected through a series resistor 491 to a summing inverting input 483 of amplifier 495. Output of amplifier 495 is conducted through ~ series resistor 490 to an inverting input 92 of n difference amplifier 494.
A non-inverting input 96 of the difference amplifier 494 is provided with a fixed reference voltage from a regulated DC voltage source 498 and an inverting amplifier 501, which are conventional circuits.
The output 504 of the difference amplifier 494 i8 connected to the analog input terminal of an analog-to-digital converter 478. The circuits involving subcircuits 494, 495, 498, and 501 are represented by the signal-conditioning circuit block 474 of Figure 36.
A/D converter 478 is a commercially available semiconductor device and may be model No. AD670 marketed by Analog Devices company of Norwood, Mass. It converts the analog signals that it receives on line 476 to 8-bit digital data at its eight output lines 480. Those lines 480 conduct the digital signal to input terminals of the mi~;Lucul~LLuller 482.
The mi~Luco-.~Lùller 482 may be of the type Intel 8051, 8751, etc. The one used in this ~nho~ L
is a Model DS5000, which is available from Dallas Semiconductor company of Dallas, Texas, and which is the same as Intel 8751 except with more internal R~N. A
crystal 510 and two capacitors 512 are co""ect~ to a ~rm;n~l of mi.LuyLucessor 482 to dletermine the clock frequency of the miuLuU~ uCeSSUL . Five lines generally indicated as 514 are connected from outputs of the microcontroller 482 to inputs of multiplexers 470, 471 to enable the miuLo~u.-~Luller to step multiplexers 470, 471, through the twenty-four sensor inputs sequentially by : .

WO 92/09941 2 0 9 ~ 4 4 2 PCr/US9l/08892 addressing them one at a time. Output lines 484 from the mi~:L~LUCess~lL lead to a ;cations chip 485 and therefrom to a communication board related to a main mi~L~ ..LL~ller. C i rations chip 485 may be a 5 Motorola ^actured and marketed NC3487.
The following table is a list of L
types and values, as used in the circuit of Figures 37A-D.
~3m~ ~3~ Value or Tvl~e 10 MC1 Capacitor 1. 0 ~f MC2 Capacitor 1. 0 ~f MC3 Capacitor 1. 0 ~f MC8 Capacitor 0.1 ~f MCg Capacitor 33 pf 15 MC10 Capacitor 33 pf MCll Capacitor 0.1 ~f MC12 Capacitor 100 uf MC13 Capacitor 0.1 ILf MC14 Capacitor 0.1 ,uf NCR1 Diode lN914 MCR2 Diode HLMP6500 MQ1 Transistor 2N2222 MR1, R2 Resistor 100K Ohms MR3 Resistor 150K "
MR4 Resistor 100K "
NR5 Resistor 1. 69K "
MR6 Resistor 2 . 21K "
NR8, 9 Resistor 2 . 2K "
MR11, 12 Resistor 10K "
30 MR13,14 Resistor 2.2K "
MR15 ,16 ,17 Resistor 100K "
MRl9 Resistor 43K "
MR20 Resistor 75 "

WO 92/09941 PCr/US91/08892 `2095~4~ --R7 Resistor 4 . 7K
R10 Resistor lOOK "
R18 Resistor 150K "
El-24 Hall Sensor 915512-2 Ul,Ug Multiplexer AD7506 MU2 Diff.Amp. LF347 MU3 Comm . Chip MC3 4 8 6 MU4 Mi-:Luco~lLL. DS500032 MU5 A/D Converter AD670KN
10 MU6 Logic Circuit 74LS132 MU7 Comm. Chip MC3487 MU8 DC Regulator LM317LZ
MY1 Crystal 12MHZ
Data Processinr~
A simplif ied algorithm is shown in the f low chart of Figure 38 to explain how the mi.;LuyLUce6sul 482 det~rmi n~C the lateral and longitudinal positions of floor-mounted magnet 6 as the array of Hall sensors 24 passes generally over the magnet 6 . PLUYL i ng techniques for accomplishing the specified steps, 6een in Figure 38 and also in Figures 39 and 40, are known in the computer art.
Inlt;~1i7in~ and UPdatinq of the Null Voltaqes When the update marker system is activated the null voltage of each sensor 437-460 is measured by multiplexing the outputs of the sensors one at a time.
The respective null signals of each of the sensors are measured several times, added together and divided to obtain an average value. Averaging is neC~cc~, y to reduce the effects of errors in mea~uL~ Ls of the null voltages . Each sensor has a dif f erent average null voltage; an average is computed for each sensor alone.
Because the sensor outputs vary with temperature the null voltage is L. -c~lred (updated) for WO 92/09941 ~ ; 4 ~ 2 PCr/US91/08892 ~11 of the sensors after each time that a magnet is ~-c.vaI~ed. This reduces errors that otherwise might result from differences in t~ atuLe along a vehicle's path .
- A simplif ied description of the program of Figure 38 6tart6 at a flow line 520. In block 522 the null voltages of the sensors 437-460 are measured. To do this the mi~LuuLUcessoL 482 of Figures 37 A-D address the first sensor by way of multiplexers 470, 471. The signal from the first sensor passes across line 472 to the difference amplifier 494 and the A/D converter 478, thence to the mic:~vu~ uces~u~ 482, Figures 37 A-D, where it is temporarily stored.
Returning to Figure 38, in block 522 the multiplexers 470, 471 are strobed to multiplex in the null voltage of the second sensor, etc. until all sensors have been measured. The entire sequence is then repeated several time6 in block 522, starting again with the f irst sensor. In block 524 all of the null readings of the first sensor are av~:Layed and in block 526 the average value of null readings of the f irst sensor is stored .
This averaging and storing process is performed for all twenty-f our of the sensors .
Detec~ion of a Maqnet After the null voltages have been stored the program goes into a wait loop 528. In the wait loop the mi. Lu~Lucessor 482 continuously polls each sensor 437-460 to ~lotP~; n~ whether or not a signal level in excess of a predetPrm;nPd threshold level exist6, which would 3 0 indicate the presence of a magnet nearby .
Details of the wait-loop are as follows. Block 530 shows the polling of sensor signals. In Block 532 the previously stored null voltage corrPcponrl;n~ to each sensor is subtracted from the signal output of that sensor to obtain a difference signal, lepLas .,~ing the Zi~L~ l of a magnetic field. In the block 534 the difference signal is te6ted to ascertain whether or not ___ _ , _ _ _ _ _ _ _ _ _ . . .

-WO 92/09941 -- 2 ~ ~ ~ 4 4 2 ~ PCr/Usgl/08892 ~

it exceeds a predet~rmin~d threshold level, which is set 50 as to dlfferentiate between noise and true magnetic marker signals. If the difference æignal is below the thre6hold leYel the wait-loop routine is repeated.
In another pref erred ~rnhoA i - L, the program rlow of which is seen in Figure 39, the averaging and storing process is continued through a wait loop 528 ' .
In this ~ , a running average of each null voltage is calculated in block 550 by the following equation:
Nj(t) = (K1 * Nj(t-l) + rj(t) ) / (K~+l) where: j L~es~.. Ls the figure number of a selected sensor ( i . e . j = 4 3 7 thru 460).
t is the time of the current sample .
t-l is the time of the previous sample .
Nj (t) is the average meaDur~ of each null voltage at time t for sensor j.
K1 is an integer multiplier which detPr-ni n~ the time or sample by samplé weighting of past and present meaDuL- Ls on the current running average voltage calculation. (K~ may be on the order of lO0. ) Nj(t-l) is the average meaDuL~ L of each null voltage at the previous sample or time t-l f or sensor j .
rj (t) is the raw voltage meaDuL . L
of the voltage at time t for 3 5 sensor j .
When a difference signal is found to exceed the WO 92/09941 2 0 9 ~ ~ 9 2 PCI/US91/08892 predet~r-n; ned threshold level, the null voltage calculation i5 terminated. All other program functions in wait-loop 528 ' are the same as those of wait-loop 528 .
Selection of a Grou~ of Sensors If the difference signal is large enough, block 536 stores the difference signal. It then finds the sensor having the greatest such difference signal and the sensor having the second greatest. The program of mi-;Lu~uCeSsol 482 identifies the two closest sensors on the left side of the sensor that has the greatest difference signal, and the two closest sensors on the right side of the sensor that have the greatest difference signal, in block 538. Thus a group of five sensors is defined. The program then refers in block 540 to a lookup table that is stored in its memory to determine the distance to the magnet from each sen60r, based on the magnitude of the signal received from the sensor .
Two tables, as shown by example below, relate the voltage measured by each sensor (437-460) to the absolute distance to the center 557 of magnet 6. Table 1 is a lookup table comprising voltages measured at inuL~ tal distances by a sensor (437-460) from a magnet

6. Table 2 is a table providing the actual distances from the sensor to the center 557 of the magnetic field as derived from currently used sensors (437-460) and magnet f ield strength .
Relative Table 1 Table 2 ~emory Location (Measured Voltage) (Radial Distance) 30 o 142 raw ADC units 0.0 inches 139 0.0941 2 133 0. 1882 3 124 0 . 2823 4 112 0. 3764 35 5 99 0 . 4705 6 85 0 . 5646

7 71 0. 6587 WO 92/09941 PCr/US91/08892 2~95442 48

8 58 0 . 7528

9 46 0 . 8469 37 0.9410 11 29 1. 0351 512 23 1. 1292 13 17 1. 2233 14 13 1.3174 9 1. 4115 18 7 1 . 5056 17 4 1 . 5997 18 3 1. 6938 19 2 1 . 7879 The step of looking up the distance from the sensor to the magnet is performed by the mi~L~Loces60r 482, and is represented by the block 540 of Figures 38 and 39. The five sDlectPd sensors are denoted by S
(where i = -2 to 2) and the center sensor or sensor having the greatest measured voltage is S0. Before a search is made to correlate each measured voltage with 20 the related distance to the center of magnetic flux, the stored null voltage, Nj, is subtracted from the currently derived raw signal from each sensor (437-460) to provide a search variable, Ej, devoid of the null offset error as shown in the following equation:
2 5 E; = S0 - Nj A sequential search through Table 1 is performed for each search variable Ej each time the group of f ive sensors is sampled. To determine the distance from each SDl ecteA
sensor (S~2 ~0l2) to the center of magnetic flux, the table 30 is searched until the di~erence between the value in Table 1 and the search variable changes sign. When the sign change occurs, the search variable i5 detPrmlnP~ to be between the last and next-to-last Table 1 value used.
An interpolation variable, I, is next calculated as 35 follows:
I = (E; - T~) / (T~ ~ -- Tl) WO 92/09941 2 Q 9 ~ 4 4 2 P~/US9l/08892 where the previously undefined variables are:
k is the relative memory position of the last Table 1 value used.
TA l epL ~6~ LS the Table 1 value at relative memory position k.
T~ ~L.~8C:11LS the Table 1 value at relative memory position k-l.
also:
R L.~sc:llLs a radial distance mea~,uL~ t.
of Table 2.
R} re~ senl_s the Table 2 value at relative memory position k.
R~ ~ represents the Table 2 value at relative memory position at k-l.
. , 15 The radial distance, Di, from each sensor to the center of rlux of magnet 6 is then calculated a6:
Di = I * ( R~ l - R~ ) + R~ ~
To calculate the position of the center of flux of magnet 6 from a common fixed point, such as array end 560, on 20 the array 24, each Di i5 treated as a lateral vector, the sign of which is detPnm; nP~ by its position relative to E;ensors having the greatest and second greatest difference signals as herebefore related. The position of the center of flux 25 of magnet 6 from the common f ixed point 560 is then calculated by adding or 6ubtracting each Dj ~Ppc-n~l i ng upon the sign of the vector to or from linear distance Li of each sensor from array end 560 as shown in the following equation:
Pi -- Li +/- Dj A further correction may be made to relate the center of rlux of magnet 6 to the centerline 164 of vehicle 2A by adding a constant which r~ s~l.Ls the distance from fixed point 560 on array 24 to centerline 164 of vehicle 35 2A. See Figure 34.

WO92/09941 ~ O ~ S 4 ~ 2 Pcr/US9l/08892 ~

Averaae Lateral Position In block 544 an average is taken of the five estimates of the location 145 of the magnet with respect to the centerline 559 of the vehicle. One estimate is 5 availabie from each of the f ive sensors of the group (having asterisks in Figure 34) whose middle one is the sensor of t~LLul-y~ signal.
In this example, sensor 445 is 50, sensor 443 is S-21 sensor 444 is Sl, 8ensor 446 is Sl, and sensor 447 is

10 S2-Arter each of the f ive sensors have beensampled, an average estimate of the position, Xt, of the center of f lux of magnet 6 is calculated as shown below:
Xt = (P2+P i+P0+PI+p2) /5 + C
where C is the distance 182 from the distance from fixed point 560 on array 24 to the centerline 164 of vehicle 2A.
The accuracy of measuLl L is further ameliorated by a running aver~ige of the successively 20 measured values of Xt. Though other equations may be used to calculate the running average, the following equation is employed in the currently pref erred r~mho~
X(t) = (K2 * X(t-1) + X(t) ) / (Ki + 1) where X(t) is the running average of the : ~ meat,uL~ of the center of f lux of magnet 6 f or the 6eries of f ive sensors measured at time t and related to the centerline 164 of vehicle 2A.
X(t-l) is the previous running average of the meait7uL~ ~ of the center of flux of magnet 6 for the series of f ive sensors measured at time t-l and related to the centerline 164 of vehicle 2A.
K2 is the f ilter or decay constant WO 92/~994l 2 0 9 ~ 4 ~ 2 PCr/US91/08892 for the running average. K2 is on the order of three in the currently preferred o~ho~
As one familiar with computer addressing would know, the 5 values of measured voltages for Table 1 need not be derived from incremental distances, but only from measurements taken at known, regularly increasing or decreasing distances which are then stored in the related memory location in Table 2. New and useful Tables 1 and 10 2 may be generated for combinations of sensors and magnets which yield different voltage versus distance values by measuring the voltage as a function of distance for the new combination. As seen in Table 2, in the above example, the radial distances stored in incremental 15 memory locations are even multiples of . 0941 inches.
Ti of Peak Sensor Siqnals The next program function, performed in block 542, is to determine whether or not the peak of sensor voltage has been passed. The peak values of output 20 voltage from the Hall sensors of array 24 occur when the ~Lrray 24 is directly over the flo~ -Led magnet 6.
When the reading of the sensors start to decline the array of sensors has passed over the center of f lux of magnet 6. This condition is detected by block 542 by 25 conventional pl~/yL in~, ' l~ccuracY of the Mea_ r L 1, The combination of precalibrating each sensor prior to meaLuL~ t to take out the offsetting null voltage and averaging and calculating a running average 30 until the peak voltage is reached provides a mea~uLI
of significantly i ov~d accuracy. The accur~cy of the lateral position mea~uL~ ~ 145 is 0. 02 inch.
OUt~ut The process of selecting a group of sensors, 35 looking up distances and averaging them is a form of cross-correlation of received signals with a stored f ield , WO 92/09941 PCr/l 3S91/08892 pattern. Thi6 re6ult is transmitted, block 546, from the mi~;Lu~Luc~n~uL 482 to a main miu,u~Luue~ur, not shown.
It is transmitted promptly when the peak readings are detected, 50 the time of tr;~ncmi Cci nn of the data serves as an indication of the time at which the sen60r array 24 crosses marker magnet 6. In this way both lateral and longitudinal position information are obtained from one passage of the array 24 over magnet 6.
Data from block 546 is transmitted to the main microproces60r board. Data f low among the mi.:Lc~.Loc,2ssors in AGV 2A are described in detail later.
The program, at point 548, then returns to the starting program flow line 520 of Pigures 38 and 39.
Another: ' ;r- t having two arrays of sensors such as array 24 is also feasible.
Ref erence is now made to Figures 4 0-4 2, wherein a second preferred: -a;r-nt is seen. In the 8econd omhnA; r L, two magnet5 6, 6 ' are placed in suf f iciently close proximity that magnetic f lux from each of magnets 6, 6' is sensed by a plurality of sensors 437-460 concur-rently, yet separation 163 of magnets 6, 6 ' is suf f icient to permit 1 n,a.~p~na.~nt processing of signals derived from each magnet 6 or 6 ' .
As seen in Figure 41, exemplary path 557 of the center of flux 557 of one magnet 6 is the same as the path described in Figure 34. A second path 657 is seen for second magnet 6 ' . The table below summarizes the results of signals derived from two CUII~ULL~ L1Y measured magnetic paths 557, 657, showing the assumed greatest 6ignal level sensed for each magnet, next highest level and sensors active for the mea_ur. L of position of each magnet (indicated by a single asterisk (*) for magnet 6 and a double asterisk (**) for magnet 6 ' ):

WO 92/09941 2 0 9 ~ ~ ~ 2 PCI/US91/08892 Relative sensor First magnet (6) Second Magnet t6 ' ) - Pgsit i f~n ~m~L . ~umber S-1 444 452*

S1 446* 454 * indicates the sensor adjacent to the 6ensor having the greatest signal magnitude and having the lo second greatest signal magnitude thereby providing an indication the center of magnetic flux (145, 645 lies therebetween.
Figures 40 and 40A-B show a simplified flow ~ chart of the logical and calculational steps for 15 -de~rm;n;n~ the position of the vehicle relative to each magnet 6, 6 ' . Figure 40 shows the orientation of Figure 4 OA relative to Figure 4 OB . Program f low line 52 0 connects the output of block 652 in Figure 40B to START
in Figure 40A. Program flow line 620 connects the "yes"
output of block 660 in Figure 40B to ~O.. llNU~; in Figure 40A. Program flow line 622 connects the "yes" output of block 654 and the "no" output of block 542 of Figure 40A
to START 2 in Figure 40B.
As before described, the null offsets are 25 calculated during a known null period as specified in blocks 522, 524, and 526. As earlier described, in Figure 39, a NAIT LOOP 528 ' provides an updating of the null calibration for each of the sensors until an over threshold measurement indicates detection of magnetic 30 flux of a first magnet 6 or 6'. Upon such detection as part of block 236 activity, the sensor values are stored and the sensor having the ~ nyt:~-L signal is selected as earlier described for block 536 in Figure 38. In addition in block 236, a fir5t sensor group active flag 35 is set to signal a first magnet position measurement is active .
..

WO 92/09941 ~ PCr/US9l/08892 209~4~2 As earlier descrlbed, the activities of blocks 538, 540, and 544 select the group of sensors used in the calculation of what is now the f irst sen60r group, interpolate the distance from each sensor of the first group to the center of --gn~tjr~ flux of the first detect~A magnet and average, then calculate a running average of the position of the vehicle relative to the magnet. Decision block 542 branches to a block 546 ' when the peak value of the f irst sensed signal is detected or to a second path headed by START 2 before the peak is disc~,veL ed .
At START 2, input program flow line 622 leads to decision block 624 wherein a decision i8 made whether or not a second group active f lag is set indicating a signal has previously been detected from a second magnet.
I~ the second group flag is not set, a single pass through blocks 630, 632, and 634 is made. Blocks 630, 632, and 634 comprise ~1OYL in~ functions which are similar to those described for blocks 530, 532, and 534, except blocks 630, 632, and 634 only process information related to sensors of array 24 not involved with the first group. If no threshold is detected in block 634, an updated null calibration is calculated for each sensor which is not part of the f irst group and a branch is made T0 CC ~ lNU~' to merge with program flow line 620. If a signal above threshold is detected, a branch is made to block 636 wherein the appropriate signal values are stored and processed as in block 536 for a second group of sensors and the second group active f lag is set .
The program proceeds directly from block 636 to block 638. If the second group active flag is set upon entry at program flow line 622, a branch is made directly to block 638 therefrom.
Sequentially, blocks 638, 640, and 644 perform the same functions upon data received from sensors of the second group as blocks 538, 540, and 544 perform upon data received from sen50rs of the first group. Decision _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ~0 92/09941 2 Q 9 5 4 ~ 2 Pcr/us9l/o8892 block 642 determines whether or not a signal peak, as before de6cribed, has been reached. If not, the process continues to decision block 660. If 60, mea6ured position values, as derived from both magnet6 6 and 6', 5 are tran6mitted to the main proce6sor for use in navigation and gu;~nce updating, the first and second group active flags are reset as shown in blocks 646 and 652. From block 652, the logic path proceeds to START at program flow line 520 to repeat the ~unction pr-~limin~ry 10 to the search for one or more additional magnets along the vehicle ' s path.
From decision block 660, a branch is made to block 638 if the first group active flag is reset indicating a peak has been detected for the first 15 measured magnetic f ield . If the f ir6t group active f lag is set, the program proceeds to program flow line 620 whereat block 538 is entered to subsequently process the output of the f irst group of sen60rs dedicated to making a mealcu~ t of the position of the first detected 20 magnetic $ield.
If within block 542 a peak voltage i6 detected, the pLOyLal.... proceed6 to block 546 ' wherein the mea6ured position detF~rm;n~ by fir6t group mealru.~ t6 are stored for later recovery and transmission to the main proce660r and the first group active flag is re6et. From block 18 ', decision block 654 is entered, wherein a branch i6 made to proceed T0 START 2 through program f low line 622 if the second group active flag is set or to proceed to block 656 if the second group active flag is re6et. At block 656, only the fir6t group measured po6ition is reported ba6ed upon only one magnetic f ield having been detected and no concurrent meac,u~ ~ L having been made.
G~ ~ire Guidance Svstems -The various types of guidance systems are used at different times for controlling the AGV 2A, each type being used under control and direction of AGVC 13 and AGV

-WO 92/09941 _ PCr/US91/OX892 2A motion control ~, ùcessor 61. Figure 4A shows a motion control proce6sor 61 of a preferred control system for a vehicle. The port drive wheel 8 is driven by a port motor 15. The port motor 15 is controlled by a port 5 motor controller 19, which receives control signals from a summing junction 177.
Inputs to the summing junction 177 include:
1. control signals at an input 124 that come from the inner guidance loop motion control data processor 61;
2 . an input from a t Ir.lnl Lel 33 that measures the speed of the port motor 15;
and 3. an input 173P from a ~rminAl-positioning-mode module 37 of vehicle navigation and guidance system, which will be described in detail below.
In a similar a~la..y~ L, a starboard wheel lO
is driven by a starboard motor 17, a starboard motor 17 is controlled by a motor controller 21 that is driven by the output of a summing junction 175, which receives speed ~ n~ from the motion control processor 61. The summing junction 175 also receives signals from a starboard t;~r~ Ler and from the t~rmin 11 -positioning-mode module 37.
The motion control plUOt:EiC.UL 61 receives "~ at an input 39 from a self-contained navigation and guidance system . The vehicle is driven in f orward and reverse directions, relative to the front of the vehicle, and is steered in accordance with the speeds of the wheels 8, 10. The actions of the drive wheels 8, 10 affect the vehicle in a manner that is re~Les~.,Led symbolically by a summing junction 41 and by a block 43 labeled "vehicle dynamics" on Figure 4A.
The spacing between the wheels and other factors are represented by the block 43. ûutputs of the block 43 are represented symbolically at a point 45. The WO 92/09941 2 ~ 9 ~ ~ ~ 2 PC~/US9l/ox892 outputs are the speed and heading of the vehicle as well as, when integrated, the position of the vehicle. The position of the vehicle controls the error signals as the vehicle moves about, for example, when it enters a ~PrminAl 9, 11. As shown in block 37 of Figure 4A the t t~rm;nAl-positioning-mode of the vehicle navigation and guidance system inrl~ c the antennz assembly 47 and an analog circuit block 49, both of which will be described in detail.
0 C -n~lc for control of the vehicle are at t~r-n;nAl 39 on the left side of Figure 4A. C nrlc and feedback signals such as 173P and 173S are conducted through the summing junctions 177, 175 to the port motor controller 19 and the starboard motor controller 21 respectively. They drive the port motor 15 and the starboard motor 17 respectively, which drive the port and starboard wheels 8, 10 respectively.
When the vehicle 2A enters a ~rminAl having a passive loop floor mat 51, it comes principally under the control of the terminal-positioning mode of the vehicle navigation and guidance system, etc. This system produces signals 118, 120 that are input to the motion control processor 61 of Figure 4A.
When a vehicle with an incoLL~ ~ lateral position (e.g., with an offset from the centerline of the tPrm;nAl), enters a t~rminAl an error signal is generated by the t~m;nA1-positioning mode of the vehicle navigation and guidance system. The signals at lines 118, 120, in combination, produce an error signal which has such polarity (see also Figure 12B) as to operate the motors 15, 17 to steer the vehicle in a direction to correct the errojr of position. Antenna output signal conditioning cixcuits similar to those seen in Figure 12B
for front-end antenna signal conditioning, not shown, are located at the rear-end of the vehicle but are of opposite hand.
When the vehicle has proceeded longitl-~; nA l l y WO 92/09941 2 0 9 ~ 4 ~ 2 PCr/US91/08892 ~

to where a wire-cross exists, another signal, on line 300 of Figure 4B, notifies the outer loop pLvce6suL 67, which takes ~ v~L iate action of altering speed n-7c . The antenna assembly 47 of ~igure 4B includes antennas that 5 are receptive to the transverse wire-crossing portion 87 of the loop 55, as will be described in more detail below in sections relating to wirL ._L v66ing positioning of the ve7licle. The longitudinal position of the vehicle is controlled by the motor controllers 19 and 21, which operate the motors 15, 17 80 as to move the vehicle forward and back as nPcPcc:~ry to position it over the wire-crossing portion 87 of the passive loop 55, etc.
ovPrview of Intt:7 ~ ,evl ions of Maior Guidewi_e 6llhcy67~prllc 15 - - Figure 4B is a simplified diagram showing the relationships between major gubgysl;ems of the tPr777in;71-positioning mode of the vehicle navigation and guidance system .
C n~7c from the AGVC 13, which stores map-20 like route and vehicle-location in:Formation, by wireless transmission to _ ; r~tions block 13 ' go to an outer loop mivLvcv..LLvller 67 whose outputs go to a motion control mi.;Lv. v..~Loller 61. They then pass through a D/A
converter 133 to a sum7jing junctio~ 175. The output of summing junction 175 goes to a controller 21 and forward/reverse block, which drives the starboard motor 17 and wheel 10. Only the starboard circuits are being described .
As the vehicle moves about to carry out the _ nr7c that it receives, feedback signals responsive to its position are generated. They are processed and entered into the control system through several rh;7nn~1 c.
A6 shown on Figure 4B, these rhAnnl~l c include a Passive Lateral Sllhrh~7nnpl at tPr777in;7l 118, a Guidewire Lateral 5l7hrh;~7nnPl at 7~Prl77;nAl 173S, a Guidewire-Crossing Sl7hrh;7nnel 261 and a Pagsive-Wire-Crossing SllhrhAnn~7, 281. The rh~7nnPl c are described briefly here to show ~092/09941 209a 442 PCI/US9l/08892 their relat i nn~h i r~, and in much greater detail in subsequent sections.
A magnetic transmitter 68 couples magnetic energy to a passive loop 55 on the floor in a t~rm;nAl.
Induced current in the passive loop 55 produces magnetic t field8 that are sensed by a receiving antenna system 47.
The receiving antenna system 4 7 compri6es separate r--gnPti r. receiving antennas for lateral positioning of the vehicle and for wire-crossing positioning of the lo vehicle.
Instead o~ being energized by the magnetic transmitter 68, the magnetic receiving antennas 47 can, alternatively, be energized by magnetic fields produced by a wire 3 in the floor, as shown on Figure 4B. The wire 3 in the floor is energized by the AGVC 13, which is represented on Figure 4B for drafting convenience by an AC generator 13A ' .
Output from the lateral-positioning system's antennas are connected to a right Lateral Channel 109, which will be described in more detail, and to a left Lateral Channel which will not be described because it is the same as the right Lateral Channel.
The right Lateral Channel divides into a Passive Lateral cllhrhAnn~l, including rectif ier 113 and an amplif ier. The Passive Lateral SuhrhAnn~l connects through an A/D converter 135 to the motion control processor 61, where it joins the command signals.
Signal6 then pass through the D/A converter 133 and are input to the summing junction 175.
Figure 4C is a simplified version of Figure 4A.
It is a functional block diagram showing elements that ~re in use when the e~uipment is in the terminal-positioning mode of operation. The Passive Lateral Sl-hrhAnnPl components, which are used in the Terminal-positioning mode of operation are s~lown. The analog cir-cuits 49 are not shown in Figure 4C because they are effectively by-pa-sed when the t~rm;nAl-positioning mode WO 92/09941 PCr/US91/08892 2~9~42 60 is operating.
The right Lateral Channel 109 also goes to a Guidewire Lateral Sl]h~hAnnPl, which starts at a shortable attenuator 180. Fig. 4B. (Most of a CULL .~ l;n~ left portion of the channel, starting at 111, is omitted from Figure 4B. ) The right Lateral Channel then goes to a b~n~lrs~fi filter 157 and other signal-processing element6.
It is switchable by a switch 170 (controlled by outer controller 67) before tPrmin 1l 173S to allow input to the summing junction 175 when guiding over a guidewire 3 in the floor, and to prevent interference from signal at 1735 when guiding over a passive wire 55 in the floor.
The wirc ~Lvssing receiving antennas are connected to a Wire-Crossing Channel at logic circuits 217, etc. These circuits produce a WiI~ ~.LU55ing signal WX and a wi~ ~ Lossing reference signal REE7iX, both of which are connected to two 51lhrh Inn~
The first of the two wi~ _Lussing sl~hrhlnnPlS
is a Passive-Wire-Crossing SllhrhAnnpl that starts with 1155-Hz b~n~lr~s filters 277 and 279. Its signal proceeds through rectif ier and logic circuits to an output terminal 281. Terminal 281 is connected to the outer loop mi-;Lucu--LLuller 67, completing a positioning-f eedback loop .
The other sllhrh~nnPl to which the Wire-Crossing-Channel is nnnnected is the Guidewire Crossing S~lh~h;lnnPl of Figure 48. It starts with 965-Hz b~n~lr~elc filters 243 and 245. The signals proceed through rectifiers and logic circuitry to tPnmin~l 261. From there the feedback signals are connPcted to the outer loop mi~Lu- u..LLuller 67, where they join the command signals from the AGVC 13, to complete a positioning-feedback loop.
When a vehicle is in a tPr~nin~l that has a 35 passive loop 55 on the floor, lateral positioning is A1 ~ hPd by means of the Lateral Channel and the Passive Lateral S-lh~-h~nnPl. Longitudinal positioning is W0 92/09941 2 ~= 9 ~ 4 ~ 2 PCr/US91/08892 accomplished through the Wire-Crossing Channel and the Passive-Wire-Crossing S~hrhAnnPl.
When a vehicle is in a terminal having an active guidewire in the floor, lateral positioning of the 5 vehicle is A~ h~d through the Lateral Channel and the Guidewire Lateral S~hrhAnnel. Longitudinal positioning is accomplished by means of the Wire-Crossing Channel and the Guidewire-Crossing S~lhrhAnn~l.
When a vehicle is not in a terminal and is on a 10 route, such as route 5, that ha6 only update magnets, guidance is accomplished by self-contained navigation and guidance .
When a vehicle is not in a terminAl and is on a route, such as route 3, in which there are actively 15 energized guidewires in the floor, lateral positioning is - accomplished by means of the Lateral Channel and the Guidewire Lateral ~llhrhAnn~l Longi~udinal positioning can be accomplished between t~rm;nAl c where there is a wire crossing by means of the Wire-Crossing Channel and the Guidewire-Crossing ~llh~hAnn~l.
MA~netiC Fiel~lc TrAnFmitt~r The subsystem 37 of Figure 4A includes a magnetic field transmitter that is shown in simplified form in Figure 5. A sinusoidal waveform oscillator 68 on the vehicle is connected through a switch 70 and an amplifier 69 to a transmitting antenna 71 to provide a magnetic field signal of frequency 1,155 Hz. The transmitting antenna 71 is part of the antenna assembly 4 7 shown on Figures 3 and 4A .
The transmitter is shown in more detail in Figure 6. The main component of its oscillator 68 is a conventional commercially available chip 68A. Its output at t~rminAl 68B is connected to the analog on-off switch 70. When the switch is in a conductive condition the 06cillator ' s signal is connected to input 69A of one side of a push-pull current driver amplifier 69.
The output at 69B of one amplif ier 69 is WO92/09941 20 ~ 5 4 4 2 ~ PCrtUS9l/08892 ~

cnnnPcted through a resistor to a point 69C, which is cnnnPcted to another pole of the analog switch 70. The output of that pole at 70B is rnnnPtecl to an inverting input 69D of another side of the push-pull driver amplif ier 69 . The output of that other side is at a tPl-TII;nAl 69E.
The output terminals 69B, 69E of the push-pull drivers 69 are connPctP~ to two series-connectP~l coils 71A, 71B of the transmitting antenna 71, as shown in Figure 6.
The analog on-off switch 70 is operated by a signal at a tPrm;nAl 70C, which comes from the outer loop mi~,Luploc~ssuL 67. The transmitter system comprising elements 68, 69, 70 and 71 is turned off by operation of the switch 70 when the vehicle is being operated in a mode in which it f ollow6 an actively energized guidewire .
The outer loop processor receives information from the AGVC 13, which keeps track of whether or not the vehicle is approaching or in a tPrTII; nA 1 .
As shown in Figure 7, the transmitting antenna 71 includes a ferrite rod 75 that serves as a core for the antenna. The relative magnetic permeability of the ferrite rod is about 2000. Mounted on the core 75 near its ends are a left-side coil of wire 77 and a right-side coil 79. The push-pull drivers 69 are connected to the coils 77, 79 with such polarity that the coils produce reinforcing magnetomotive force (of the same phase~ in the f errite rod 75 .
The lateral position of the transmitting antenna 71 relative to the center 81 of the floor loop ~ssembly 54 has very little effect on the amount of current induced in the passive loop 54 within a wide lateral range between the transmitting coils 77, 79 because the amount of magnetic f lux linking the loop 54 3 5 does not change appreciably within that range . The electric current induced in the loop 54 is, however, inversely dPrPn~lpnt upon the vertical and longitudinal WO 92/09941 2 0 9 ~ 4 ~ 2 PCr/US91/08892 distance between the transmitting antenna 71 and the central wire portion 81 of the loop 54.
The operation of the tranEmitter i6 as follows:
The oscillator 68 produces a signal which can be 5 connected through the analog on-off switch 70 to the push-pull drivers 69. The output signal from the push-pull drivers 69 energizes the transmitting antenna 71.
The transmitting antenna produces a magnetic rield that extends downward to encircle the wire element 81 of the loop 54 (or any wire that i6 within the range of the transmitting antenna , e . g ., a guidewire in the floor). In the case of a loop such as loop 54, the AC
magnetic field produced by the antenna 71 induces a current in the wire segment 81, and that current produces 15 a magnetic field ~uLluu--ding the wire segments 81, 87, etc. of the loop 54.
Receivinq AntpnnA~ An~ r~linq with Wires on Floor Figure 7 also shows a receiving antenna as6embly 91. It detects magnetic fields produced by 20 currents in wires on the floor. In this preferred ~ o~ L, a single ferrite rod core 93 is used, with one receiving coil 95 mounted near the left end ûf the rod and another receiving coil 97 mounted near the right end of the rod 93. Alternatively, two shorter ferrite 25 rods can be employed with a f ixed lateral space between them, each encircled by only one of the two receiving coils 95, 97.
In this '-: ~ i - L the receiving antenna assembly 91 is mounted parallel to and close to the 30 transmitting antenna 71. The signal that the receiving antenna assembly 91 receives has two _ ~n~nts: (a) a signal from either the passive loop of wire 54 or a guidewire in the f loor and (b) a direct signal from the transmitting antenna 71 if it is on. Because the 35 position of the transmitting antenna 71 is fixed in relation to the receiving antenna 91 the undesired direct signal -n~nt is relatively constant, so it can be WO 92/09941 PCr/US91/08892 ~
20~5442 64 ~l~rlllrt~
Referring now to the -nt of signal received from the wires in the flsor such as the wire 81 of the loop 54, the current in each coil 95, 97 of the 5 receiving antenna assembly 91 depends upon the nearnes6 of the receiving antenna 91 as a whole to the plastic floor mat 51 and upon the lateral displacement of the receiving antenna 91 from the center wire segment 81 of the pa66ive loop 54.
The relationship between received signals and lateral displa~ t is relatively linear for the central 90% of the lateral ~ t~nre between the two receiving coils 95, 97, Figure 7. The ferrite rsd 97 helps to provide this linearity. Figures 8 and 9 illustrate the manner in which magnetic f lux produced by electric current in the wire 81 tFigure 3) enters the ferrite receiving rod 93 and links the coils 95, 97. In order to facilitate the explanation, Figures 8 and 9 are not drawn to scale.
In Figure 8 the receiving antenna 91 is centered laterally over the current-carrying conductor 81, while in Figure 9 the antenna 91 is offset laterally from the conductor 81. The direction of lines of magnetic f lux is shown by a stylized line sketch 96A in Figure 8. Other lines of flux 96B, 96C of course enter the ferrite rod at its left- and right-hand ends, and hence encircle the turns of the coils 95, 97. The Figure 81 can ~ sel.~ several turns of wire in some l~--` o~ Ls.
In Figure 9 the flux line 96A encircles the coil 95, because of the offset position of the antenna 91. The flux line 96B stiil enters the left end of the rod 93 and encircles the coil 95. When the vehicle is offset, the partially shown flux line 96C no longer encircles the right coil 97. This ar r cllly. L, in which a single ferrite rod is used for both receiving coils, has been f ound to improve the linearity of the induced WO 92/0~7941 2 ~ 9 ~ ~ 4 2 PCr~US91/08892 signal in the receiving system as a function of the offset of the vehicle from the current-carrying ~ ~n~7~7~ t~r 81 .
A graph of the amplitudes of signals induced in 5 - the receiving antenna coils 95, 97 is shown in Figure lOA. The abscissa 137 r~ 8 tle lateral offset of the vehicle from the longitudinal centerline of the 7'~7~7;nAl 11. The ordinate 143 of the graph of Figure lOA
represents signal strength at the coils 95, 97.
In particular the starboard receiving antenna coil 97 E,,uduues a 6ignal shown by a curve 139, and the left antenna receiving coil produces a signal shown by a curve 141. h7hen the vehicle i5 exactly in the position defined by the ~roy~, ~ lateral offset and represented by the vertical line 143 of Figure lOA, the signals 139 and 141 cause the wheels 8, 10 to rotate at equal speeds.
For example, when the lateral offset is zero and when the vehicle comprises an of f set such as at the vertical line 145, the left antenna 95 receives a much ~LLUIIge1 signal, as indicated by a point 147 on the curve 141, than does the right antenna coil 97, as indicated by the weaker signal at a point 149 of the curve 139. The result is that the left wheel 8 is then driven slower than the right wheel 10 and the vehicle ' 8 position is corrected to center the vehicle over the guidewire as it moves forward into the 7-~7~7inAl 11 or, alternatively, along a guidewire in the floor in the t~7~ni7~Al 9.
From time to time, it may be desirable to drive the vehicle with an offset lateral to a guidewire. This i8 accomplished under program control by the motion control ~.~cessuL 61 wherein a lateral offset bias is digitally added to one of the signals over 7~7~777;nAl~ 118, 120, after digitization. As seen in Figure lOB, which comprises the same axes and curves seen in Figure lOA, a desired offset 145' away from center line 143 estAhlic~7e~
two curve 139 and 141 intersections, 149 ' and 147 ', respectively. The lateral offset bias is calculated as WO 92/09941 2 0 5~ 4 2 PCr/US91/08872 ~
the difference between the values at intersections 149 ' and 147 ' and comprising a sign opposite an error on the same side of center line 143.
HYbriditv of Self-Con~Aln~l Naviqation-and-S t~ Anr~- and Prol~Ortional-positi~n~n~ System - Figures relating to hybridity include Figures 4 ~nd 13 . The vehicle navigation and gui ~lAnre system, in the sQlf-contained mode, operates by 6tarting with a known position and heading and measuring the distances traveled by both the left and right sides of the vehicle.
It integrates those distances to keep track of the location of the vehicle. The position is updated per~o-l1cA1ly by ~tPrtin~ a magnet of known position such as magnet 6 in the f loor over which the vehicle travels .
The AGVC 13 keeps track of the status and position of each vehicle. The AGVC 13 has terminal inf ormation and a map of the path layout stored in memory. When a vehicle is directéd to a t~rm;nAl, such as terminal 11, that has a passive floor loop 54 and not an active guidewire, the AGVC 13 tells the outer loop es~L 67 to guide in the t~minAl-positioning mode of the vehicle navigation and g~ Anre system. r n~_ and other signals pass between computer 67 and computer 61 on a line 67A of Figure 3. The outer loop guidance mi~ c~,s.~L.,ller 67 then sends a control signal on a line 187 (Figures 6 and 15) to a switch 70 that energizes the transmitting antenna 71. It also ~ends a control signal to another switch 185 that causes attenuation of the guidewire-signal channel (t~rminA1c 153 and 155) of Figure 15 (and Figures 14, 16A).
The active guidewire-signal channel ' s error signal at t~rTninA1 169 of Figure 16B is switched off so that it does not interfere with t~e passive wire loop's signal at t~rmi nAl c 122 and 124 . This insures that the passive wire loop's signal (Figures 12 and 13) completely controls the vehicle. Nore detailed descriptions of the circuits involved are presented below.

~O 92/09941 ~ 2 Q 9 ~ 1 4 2 Pcr/US91/0~892 Lateral Positinn i nn of a Vehicle at a Terminal Havi nrr a Passive Floor Loo~
Figure 11 shows a conductive loop that is - short-circuited to itself and doubled over so that it has S two turns. One, two or any other convenient number of turns can be used. If preferred, separate superimposed shorted loops could of course be u6ed instead. They are folded to form the skewed figure eight of Figure 11 in order to produce a wire cross at any desired position.
Loops can of course be used for precise positioning of vehicles at places other than tPrmin~lc if desired.
The location of an automatic guided vehicle 2A
i8 shown and its antenna a6sembly 47 is indicated on the vehicle. The longitudinal conductors are designated by the reference number 81 and the transverse or cross wires are designated 87.
Figures 12A and 12B show a circuit diagram of a portion of the receiving equipment for receiving magnetic field information. The equipment of Figures 12A and 12B
is part of block 151 of Figure 14. In Figure 12A the receiving antenna ' s coils 95 and 97 are shown at the left side of the figure with one terminal of each coil ~nnnPctPd to ground. The instantaneous polarity of one coil relative to the other is indicated by the dots.
The circuits of Figures 12A and 12B are Dy LLical for left and right signals so only the right =
channel will be described in detail. Coil 97 is connected to a preamplifier 109, which serves also as a lowpass filter to ~u~ 5S high-frequency noise. The output of the preamplifier 109 is connected to a b;ln~lr~cc filter 109A with center LL~U~11CY equal to the frequency of the transmitting oscillator 68. The output of the 1L~C filter is rectified by rectifier 113 to convert the signal to a DC value.
The DC output of rectifier 113 is connected via tPrmin~l 113A to a shifting amplifier 117. The non-inverting input of that same amplifier receives a bias ,, _ _ _ . ;

2~V9~4~2 fro~ an adjustable voltage-dividing biasing circuit 129A, which, at the output of amplifier 117, offsets the signal that was received from rectifier 113.
The bias of amplifier 117 is a DC bias for 5 offsetting the direct magnetic coupling received from the transmit antenna. The purpose of the bias is to remove as much of the direct coupling _ - -t of the signal as possible 60 that only the signal from the guidewire is amplified, thus Pn~hl;ntJ a subsequent analog-to-digital 10 converter 135 to be a high-resolution type.
It would not be necp~c~ry for the bias 129A to be adjustable because it is sufficient to offset the signal only approximately, but it is adjustable in the preferred embodiment. The left signal is later 15 subtracted from the right signal in the motion control processor 61 anyway, so the portion of the direct signal that is not properly biased at amplif ier 117 would be t~nrel Pd by the subtraction if the antennas are centered with respect to each other. However, an adjustable bias on both right (129A) and left (129B) sides eliminates the need to ad~ust the antenna zlssembly, and allows bias adjustments to be made manually any time after the antennas are fixed in position. An automatic bias adjustment . `--';r-~t is described below in a section called Automatic Bias-Setting Embodiment.
The motion control processor 61 can also observe what the offset is when the vehicle is far removed from any floor wire, store that offset value, and use it to ~ te the signals received while 30 processing.
An inverting amplifier 131 receives the DC
output signal from the amplif ier 117, and a half -wave rectifying, unity gain amplifier LL4, which follows amplifier 131, outputs values greater than or equal to 35 zero as required by the A/D converter.
In a similar manner the left-coil signal ~rom .

WO 92/0994l 2 0 9 5 4 ~ 2 PCI/US91/08892 coil 95 i5 p~OCeBE~d by circuit ~ rst~ 111, 115, 119, 132, and 116, to provide another output signal, at a t~1~m; n;~ 1 12 0 .
The tDnmin~s~ 118, 120, which have DC signals 5 received from the right-side and the left-side coils 97, s 95 respectively of the front-end receiving antenna 91, are shown also on Figure 13. Also seen in Figure 14 are terminals 118 ', 120 ', which comprise DC signals received from the starboard and port side, respectively, from coils similar to coils 97, 95, but located at the rear of the vehicle. Two additional left and right sensing antenna 91 signals are routed through hAn~lr~c filters 163, 157, respectively, and thcrefrom to rectifiers 165, 159. TDr~S;n5~ 167, 160 from recti~iers 165, 159, 15 respectively, connect to summing amplifier 161, as earlier described. In addition, signals through tDnm;n~l~ 167, 160 are transmitted to A/D converter 135 through scaling resistors 167 ', 160 ' for A/D conversion and transmitted therefrom to motion control processor 61.
20 In the currently preferred DhO~S;- L, antenna 91 signals are processed directly by motion control processor 61, thereby bypassing lead-lag compensator 171. All six such inputs are connected to a multiplexed analog-to-digital (A/D) converter 135, which alternately converts signals 25 on all input lines to eight-bit digital signals at an output bus 13 6 .
Those digital signals are conducted to the vehicle ' s motion control processor 61. It is a Model DS5000 microprocessor manufactured by the Dallas Semicon-3 0 ductor Corporation .
Another input to the motion control processor61 is received from an outer loop microprocessor 67, which is an Intel Corporation Model 80186 device. The AGVC 13, ; rateS with the outer loop processor 67 .
3 5 Data is transmitted between the AGVC 13 and the outer loop processor 67 by guidewires in the f loor or by a radio link using an antenna 15.
_ _ WO 92/09941 ~ o 9 5 4 4 2 PCr/US91/08892 c ~- sent from the outer loop pL~ 550' 67 to the motion control E" o~.essuL 61 include the desired vehicle speed and the ratio of the left and right wheel speeds, which controls the radius of ;UlVO~U-~: of travel.
However, when the t~rm;nAl-positioning mode of the vehicle navigation and gl~ ~ rl~r~re sy6tem is being used the ratio of the left and right wheel speeds is 1. 0 . The speed command is the same to the lef t wheel as to the right wheel; corrective signals are generated from the receive antenna and are ~- i nr~rl with the speed n to f orce the vehicle to track the wire . Theref ore, the vehicle follows the path of the guidewire regardless of the path's layout (e.g., a non-straight path).
Nicrocomputer pl~JyLc~l..8 for speed control of wheels of 15 automatic guided vehicles are well known in the prior art. In the currently preferred ~ calculations have been simplified by ARsllm;nrl the error offset L~:~J1 s~:l.Ls the current guidewire position relative to vehicle 2A. The program which performs the calculations 20 i5 provided in detail in sur~w~.le listings.
In one travel direction, the port and starboard wheels delineate left and right direction, as is true when the vehicle is traveling in the forward direction.
However, when the vehicle is traveling in the rearward 25 direction, the port and starboard wheels delineate opposite hand directions, right and left, respectively.
For this reason, inputs 118 and 120 as seen in Figure 13 are received from the starboard and port side of the vehicle and are processed as right and left dilection 30 signals, respectively.
Digital data from the motion control pLOces~L
61 is conducted to a digital-to-analog (D/A) converter block 133. The block 133 contains two D/A converters 133A and 1338 for starboard and port signals respec-35 tively. The analog signal at each of their outputterm~n~l~ 122, 124 is cnn~ ctecl th~ough a summing junction 175, 177 to a motor controller 21, 19, to motors .~

-WO 92tO9941 2 0 ~ 5 4 ~ 2 PCr/US9i/~8892 17, 15, and the drive wheels 10, 8. See Figure 14.
During operation of the vehicle at place6 away from a t~rmi nAl the AGVC 13 and the outer loop ~JLVI.;t:S
67 provide . '- to the motion control ~LUC~ol 61, which supplies signals through the D/As 133A, 133B to control the motion of the vehicle via its controllers, motors, and drive wheels.
During operation in a ~rminAl the Ant~nnAc 97, 95 receive induced signals from a loop of wire 54 on the floor, and provide signals through the circuits of Figures 12A and 12B and the A/D converter 135 of Figure 13, then through the motion control ~Luce6sol 61, t~r-n;nAl~ 122, 124, junctions 175, 177, controllers 21, 19 (Figure 14 ) and motors 17, 15 . These error signals alter the speed _ n~C: of their respective wheels to position the vehicle laterally as desired in the t~rminAl .
S of PA~ive Loo~ Positioninq O~eration To summarize, the terminal-positioning mode of the vehicle navigation and guidance auyaLe.l_us guides the vehicle over a passive wire as follows:
First, miuLupLucessu~- 61 and 67 receive a signal from the AGVC 13 notifying them that the vehicle 2 is entering a t~rminAl such as t~rminAl 11. The transmitting antenna 71 is turned on by means of the analog switch 70, Figure 6, which is controlled by the mi~;LU~Ll~CeS~UL 61.
Signals from the receiving antennas 91 are preamplif ied. The right-coil and left-coil signals are conditioned with identical electronic circuits, so the following description covers only l:he right-coil signal.
The right-coil signal is routed through two different paths, namely the circuits of t~rminAl~ 118 and 155, Figure 14.
Within block 151 of Figuze 14, the right-coil signal is routed to a bAn-9r~ fil~er, rectified, inverted and added to (i.e., offse~ by) a bias, and .

W092/09941 ~9~442 PCr/US9l/08892~

amplified to obtain the signal at tprm;n:~l 118. It is also routed to an attenuator to obt~ in the signal at tPrm; n;~ 1 155 .
The signal at 118 goeg through a path ; n~ lrl; n~
5 the motion control pl.~C~:560. 61, (and nPc~ A/D and D/A converters), Figure 13. The signal at tprm;n5~l 155 is amplifled in a hAn~lr~C~ filter 157 and then rectified (159), and no bias is removed, leaving the difference at tPrm;n~l 160 very small. C~n~PT~Pntly the error signal 10 is very small. The ~ignal at 169 is switched off by the outer loop ~ e&sur 67 while the vehicle is traveling in over a passive guidewire, to eliminate any possible undesirable effects~ (See switch 170, ~igures 4B and 16B).
Lateral Positionina of Vehicle over Active G~ Pwires at Term;n~l~ and ~l~PwhPre In the case of tPrm;n~l~ such as terminal 9 of Figure 1 that are approached on routes such as routes 3 of Figure 1 (which have guidewires PmhP~lP~I in the floor), guidewires are uged in the floor of the tPrm;
also, to position the vehicle within the tPrm;n~l.
Figure 14 shows receiving equipment on the vehicle for guidewire operation both inside an~ outside a fPrm;n;~l, so far as lateral positioning of the vehicle is c-ulc~rl,ed.
As shown in Figure 14, guide signals from a wire in the floor enter (at tPrm;nAl~ 150) a block labeled "Antenna and Preconditioning Circuits" 151.
Portions of this block 151 were already descri~ed in connection with Figures 12A and 12B, where tPrmin~l~ 118 and 120 are shown. Other portions of the block 151 will be described ,.u~ u~ ly in connection with Figure 15, but for pu~oses of PYrl~ininq the general concept it is helpful to finish describing the block diagram of Figure 14 first.
The Antenna and Preconditioning Circuits block 151 outputs an AC signal at a tPrmih~l 155, which goes to W0 92/09941 - 2 ~ 9 5 ~ 4 2 PCI`/US91/08892 a hRn~lrAcfi ~ilter 157. This filter is tunable to either guidewire frequency, specifically 965 Hz or 1155 ~z. Two guidewire frequencies are available to enable the vehicle to select either one of two guidewire paths 5 at a f ork .
The outer loop ~JL o~ ~:8~iOL 67 alternates the center frequency of this b~n~lrAc~ filter 157 by means of an analog switch, which switches appropriate resistor values into the circuit to select the desired frequency, 10 until a significant amplitude is detected, signifying acquisition of the guidewire. The filtered signal is fullwave rect;fi~cA~ in a block 159. The result at t~A~ninAl 160, which is from starboard signal channel, is sent to a non-inverting input of a summing junction 161.
= A port channel output from the block 151 is at f~m;nA1 153. It is passed through a b~n~lrAcc filter 163, then through a fullwave rectif ier 165 . At a ttArminAl 167 it is entered into an inverting input of the summer l61. The output of the summer 161, at t~nmin 20 169, is an error signal. That error signal is passed through a lead-lag tor 171, which is tailored to the dynamics of the system as a whole to provide stability, fast response, and high accuracy.
The output of ~the lead-lag -ator 171 is 25 inverte~d and added to the starboarc~ speed command 122 from the D/A 133A of Figure 13 at ~ummer 175. See also Figure 4A for a broader view. The summer 175 outputs a signal at a t~rm;nAl 201, which is connected to the starboard motor controller 21. That motor controller 30 controls the motor 17 which drives the wheel 10, as r de8cribed earlier.
The output from the lead-lag ~ -~tor 171 i5 c~ cl also to another summer 177 without being inverted first. Summer 177 adds the ~ t.ed error signal 171 to the port speed command 124. The summer 177 outputs a signal to the port motor controller 19, which drives the port motor 15, hence the wheel 8. The P!_ --WO 92/09941 ~ 2 U 9 ~ 4 ~ 2 . Pcr/US9l/08892 ~

elements lS7 through 177 are on an analog circuit board.
Details of the lateral-control circuits on the vehicle for a guidewire mode of operation are shown in Figures 15, 16 and 17, which will now be described.
Figure 15 shows connections 110, 114 ' from the prQamplif iers 10g, 111 that were shown on Figures 12A .
The signal from preamplifier 109 goes to an attenuator 180 consisting of resistors 179, 181, and an amplifier 183 .
That attenuator is ~LLa.ll~d 50 that it can be short-circuited by an analog switch 185 upon receipt of a control signal (at a switch tPrmin~l 187) from the outer loop mi-;,u~ o cessor 67. A short-circuiting conductor 189 is connected around the attenuator 180. One output of the analog switch 185, which is a ~ouble-pole double-throw selector switch, is at a terminal 153, for the port side signal .
In an identical way, the output of preamplifier 111 goes to a switchable attenuator 193 and through the analog switch 185 to an output tPrm;n~l 155 for the starboard side.
In Figures 6 and 15 the analog switches 70 and 185 are arranged such that when the oscillator 168 is ~;cconnPcted from the transmit antenna 71, the attenuators 180, 193 are short-circuited and do not attenuate. This situation occurs when the vehicle is relying on active guidewires for gui~;~nre.
At other times, the os~ tnr 68 feeds the transmitting antenna 71 (via switch 70) and the attenuators 180 and 193 are permitted (by switch 185) to attenuate the signals received fro~ antenna coils 97 and 95. This situation occurs when the vehicle is relying on passive guidewires for guidance.
TPrm;n~l~ 153 and 155 are at the left of Figure 35 16A, which shows a middle portion of analog circuits for receiving and proc~ccin~ signals when operating in the guidewire mode. The starboard signal at terminal 155 of WO 92/09941 2 ~ ~ 5 4 ~ 2 P'~/US9l/08892 Figure 16A i8 conducted through a ~witch to a h~ntlr~s amplifier filter 157, which is tuned to one of the guidewire ~ nioS, i.e., 965 Hz or 1155 Hz. The ^- output of b~n~lr~ ilter 157 is rectified in rectifier 159, smoothed in f ilter 158 and sent to a summing s junction 161.
At the same time the signal 153 of Figure 16A
passes through a hAnllr~ce: filter 163, through a rectifier 165 and an amplifier 166, and is c~nocted to another input tormln5~1 167 of the summer 161. The output of summer 161, at tormin~l 169, passes through the lead-lag -~tor 171 to the torminAl 173.
In Figures 17A and 17B circuits are shown that follow Figure 16B and are output portions of an analog board. These output portions sum tl~e n-l~ at tormin~l ~ 122 and 124 from the mi~ L~ucessu~ 61, with the ted error signal at tDnnin il 173 that drives the motor controllers. A signal of Figure 17A at tormin~l 173 splits into tormin~ 173S and 173P. The starboard signal at 173S is inverted in device 197 and summed with the starboard speed command 122 at summer 175, then passes through some circuits 199 merely to select a forward or reverse direction of motion. It flows to an output tQr-ninAl 201 that goes to the starboard motor controller 21. The circuits of this figure are of a conventional nature 80 their details are omitted from this description, although they are shown in detail in the included drawings.
The signal at tormin~l 173P of Figure 17A is not inverted but is connected directly to a summer, 177, and passes through circuits similar to those just described to send a signal, at a tormin;~l 203, to the port motor controller 19, as shown on Figures 14, 17A, and 17B. Jl~nrti~n~S between Figures 17A and 17B are designated 174A, 174B, 174C, 176A, 176B, and 176C.
~ .

WO 92/09941 PCr/US91iO8892 209~442 76 o~eration of the Motion CQntrol PLuces~cI
The following e~uations describe the operation of the miuLu~LUcecsor 61. The speed ~ n~lc C,(n) and Cp (n) are signals that originate from the AGVC 13 and that 5 are sent $rom the outer loop mi~:lu~Lvcessor 67 to the motion control ~L~aessur 61. These signals are added in mi~;Lu~Iv~e~soL 61 to the __ ted error signal ec(n) to yield the resultant signals R,(n) and Rp(n), which serve as inputs to the summers 175 and 177, at terminals 122 and 124 of Figure 14.
The quantity e(n) is a measure of how far the vehicle is off-center from the floor wire; a zero value of e (n) means that the vehicle is centered over the wire .
The e(n) signal could be ~LUyL 'd to call for an 15 offset. If the floor wire were at an il~UULLeU~ position laterally, the fault could be - cnted by having the program cause the vehicle to operate off to one side of the wire. For example, the vehicle could be offset by two inches by simply adding a term to the error signal 20 e (n) .
The term ec(n), which is the ~ cated error signal, is the output of a digital f ilter in mi~;L~"urùce6sor 61 that provides dynamic loop _ 3tion of the closed control loop. It involves the current 25 value and recent values of the error signal e(n), as well as recent values of the ~ -~ted error signal ec(n).
rv of Guidewire Tr~ k i nn To summarize, the t~ n~l-positioning mode of the vehicle navigation and guidance system dpy~L - ~us 30 guides the vehicle on a guidewire portion 3 of an installation in the f ollowing manner . The transmitter assembly 68, 69, 71 is turned off IDY means of the switch 70 of Figures 5 and 6. Signals from guidewires, received at the receiving antenna 91, are preamplified (Figure 15) 35 and routed directly to an analog circuit board (Figure 14). The starboard and port signals C,(n) and Cp(n) above replicate, with opposite signs, the, c being ,, WO92/09941 ~ ~ 8 ~ ~ 4 2 - j PCI/US91/~8892 , .. .

received at t~rm;n~l~ 122 and 124 from the microprocessor 61. The summers 175 and 177 output speed ~ nrle:, varied slightly by error signals, to control the motors 15=and 17 to drive the vehicle.
Use sf thP Vehicle Naviqation and G~ n~ e A~aratus in Two rll;rl~nrr~ Modes -- N -lY Active Guidewire and Self Con~; n~rl NaYiqation and Guidance Certain components are u6ed in common, at tr~rm;nll~ and elsewhere, by both the terminal-positioning mode of the vehicle navigation and guidance 6ystem for pa6sive f loor loops and the guidewire guidance mode . The guidance system as a whole may have a portion of its routes (routes 3 ) in which vehicle guidance is provided by guidewires in the floor. The terminal-positioning mode of the vehicle navigation and guidance system can be used to track those f loor guidewires .
The _ -ntS that are used in common include the receiving antennas 47, the Figure 17 portion of the analog board 49, the preamplifiers shown in Figure 12A, the controllers 19, 21 of Figures 4 and 14, the motors 15 and 17 of Figure 4A, and of course the wheels 8, lo.
W; re-~ross; nrY Detection for Lonqitudinal Pos~;tioninq of Vehicles Longitudinal positioning of the vehicle 2A at termirlal 9 or 11 is accomplished by sensing the location of the vehicle with respect to a wire that extends transversely across the f loor in the terminal area .
Current in the transversely-fl; ~posr~rl conductor produces an alternating magnetic field ,-uLLvullding it. The current can be due to active conductive energization of the wire or can be induced by transformer action from a transmitting antenna on the vehicle that generates a magnetic f ield. The magnetic f ield encircles the wire so that, at a particular instant, its direction is upward at one side of the wire, is horizontal directly over the wire, and is downward on the other side of the wire.
Magnetic coils for sensing the presence and - -W092/0994~ ` ~ 0 9 5 ~ a location of the wire crossing are shown on Figure 7. The three coils on the left side are a front coil 205, a middle coil 207, and a rear coil 209. The coils on the starboard side are: front 211, middle 213, and rear 215.
When these coils are in place on the vehicle their axes are vertical so that their turns are horizontal. C~ e~tu~ ly when the middle coil 207 is directly over a current-carrying wire at the f loor, magnetic flux passes through the front coil 205 in one direction, say upward, at the same time that magnetic flux passes through the rear coil 209 in the opposite direction , i . e ., I' _L d . At that same time f lux in the coil 207 does not link any turns because the flux there is horizontal and the coil ' s turns are horizontal.
When the coil 207 i8 directly over the current-carrying floor wire, an alternating magnetic flux would therefore produce one phase of signal in the coil 205, an opposite phase of signal in the coil 209, and zero signal in the coil 207. The principle of operation of the ~ dL~l_u5 in detecting the longitudinal location of the vehicle by means of wir~ cLossing ~etection is based on these three signals.
The method of combining the three signals is shown in Figure 18, which is simplif ied in order to illustrate the concept6. Signals can occur in either coil 209 alone, or 215 alone, or both simultaneously. A
signal from coils 205 and 211 is added to a signal from coils 209 and 215 at a summer 217. The sum is inverted and added to a signal from the middle coils 207 and 213 at a summer 219. The output of summer 219 is inverted and applied as an input 220 to a NAND gate 221.
The signal from coils 209 and 215 is also inverted in an inverter 223 and is input to a summer 225.
This signal i5 added to a signal from the coils 205 and 211 by summer 225. The output of the summer 225 is inverted and applied to a second input 227 of the NAND
g2te 221.

WO 92/09941 2 0 g S 4 ~ 2 PCI/US91/08892 The signal at the first input 220 is a "wire-crossing signal" WX while the signal at t~rmin5~1 227 is a ~Wil~ Dsing reference signal" REFWX. Absolute values of the signal WX and the signal RE~WX are used at the torm~n~lc 220 and 227. The output of NAND gate 221 is tF~rm l ni~ 1 22 9 .
When a vehicle drives into a terminal it ~pproaches a ~L~...,v~:lsely lying wire 87 on the floor across the path of the vehicle. Only the left-hand coils 10 will be tl;cr~ ced. Before the vehicle arrives at the wire, all three of the coils 205, 207 and 209 are linked by some alternating magnetic flux from the wire and all three of their signals are in phase. For simplicity of discussion, this phase is referred to as "downward" flux.
When the vehicle has advanced to where only the front coil 205 has crossed the wire on the floor, the coil 205 has "upward" flux and the coils 207 and 209 still have downward flux. That is, the instantaneous polarity of the output signal froD the front coil 205 is opposite the polarity of the middle and rear coils 207, 209. When the vehicle has advanced to where the middle coil 207 is directly over the floor wire, coil 205 has upward flux, coil 207 has zero linking flux (because the flux i6 parallel to the plane of its coils), and the rear coil 209 has downward flux.
The signal at point 220 of ~igure 18 is the rear coil ' s signal plus the front coil ' s signal minus the middle coil ' s signal . When the middle coil 207 is directly over the floor wire 87 the signal from the front coil 205 is equal and opposite to the signal from the rear coil 209 so those terms cancel. At the same time the signal from the middle coil 207 is a minimum, so the signal at point 220 is zero. This LC:~LC:6~ 5 a wire-crossing position.
At that time the reference signal at a point 227 is a maximum because that signal is the rear coil ' s signal minus the front coil's signal. Since the signals ~ . ~

WO 92/09941 2 Q 9 ~ ~ ~ 2 PCr/US91/08892~

rrom these two coils 205 and 209 are of opposite polarity at that tlme, their algebraic difference becomes the sum Or the magnitudes of the two, 80 it is a maximum.
The logic circuit involving NAND gate 221 and circuits leading up to it are arrange~ 80 that when the ~ignal at 220 is crossing zero and the signal at 227 is r relatively qreat (although not no~o~rily a maximum) the NAND gate 221 outputs a logic signal at the point 229 that is suitable for indicating that the vehicle is directly over the wire crossing. Tllat output at 229 is low when a wire crossing is detected.
Details of the wirc ~Lu5~ ing circuits are shown in Figures 19 and 20, and some waveforms at selected points in the circuit are shown in Figures 23 through 27.
In Figure 19 the coils 209 and 215 are in s~ries and are ~-onnortod through a resistor 231 to one input of nn inverting summer 217. Coils 205 and 211 are s~nnQstPd in series, and are connected through resistor 233 to a second input of the summer 217. The output of the summer 217 is conno~toA through a resistor 220 to one input of another summer 219. A second input to the aser 219 comes from a series connection of the middle coils 207 and 213, through a resistor 222. The inverted output of summer 219 is at a terminal 235, which is shown in both Figure 19 and Figure 20.
The output of coils 209 and 215 of Figure 19 is connected also with an inverter 223, whose output is connected through a resistor 237 to an input of the summer amplifier 225. Another input of the summer amplifier 225 comes from the series cu--~.e~ -ed coils 205 and 211, through a suDming resistor 239. The summer amplifier 225 is connected so as to invert the sum signal .
The output of inverting ampli~ier 2 2 5 is at tormin~ 241, which is shown on both Figure 19 and Figure 20. The signal at tormin~l 235 is the wire-crossing signal itself and that at 241 is the reference wire-WO 92/09941 2 0 9 S ~ 4 2 Pcr/us9l/08892 crossing signal. The circuits of Figure l9 are used in common to detect wire crossings that are (a) directly energized as in torm; nA1 9 of guidewire routes 3, and (b) ~- passive induction loops as at tPrminAl ll.
On Figure 20 the signals at terminals 235 and 241 are connected through switching to bAn~r~cc filters 243 and 245. They are tuned to receive 965 Hz, which is the active guidewire freguency. A similar other subcircuit, of Figure 22, to be described later, is tuned to 1155 Hz, which is the frequency of the transmitter on the vehicle that is used f or exciting passive loops in the floor mat at a torminAl. The 1155 Hz circuit is connected at tormin;llc 242 and 244.
The two frequencies 965 ~z of Figure 20 and 1155 Hz of Figure 22 are used in a guidewire system for causing the vehicle to branch to a f irst or second route at a junction such as a "T", by applying an appropriate r~ uel.~;y to the guidewire when the vehicle approaches the junction. However, in a torminAl having a passive loop, the receiver cllhrhAnno1 of 1155 Hz freguency is used for detecting a passive loop signal, whose energy originated with the onboard transmitter 68, and the receiver sllhrhAnnol of 965 Hz frequency is used for detecting a conductively energized active guidewire crosswire at the torm;nAl.
Thus the 1155 Hz passive-wire-crossing sllhrhAnnol 277 (see Figure 4), is used for detecting a passive loop when the vehicle is in a t-orm;n~l ~ and is used for detecting a junction guidewire when the vehicle is not in a torm;nAl. The 965 Hz guidewire-crossing sllhrhAnnol 243 of Figure 4 is dedicated to only guidewire sensing, both in and out of torm;n~1c.
On Figure 20, the signal of torm;nA1 235 passes through switching to a bAn~lrACC f ilter 245 . Figure 21 is a continuation, at torm;nAlc 246 and 248, of Figure 20.
The output of filter 245 passes through an amplifier circuit 247, a switch 249, and an inverting amplif ier WO 92/09941 ` PCr/US91/08892 `2 0-9~ 4~

251. The output of inverter 251 is shown in the graph of Figure 23. That graph i5 the detected wi~. cLùs~ing signal at a tPrmin~ 253. ~ -That signal passes through an amplifier 255 5 that eliminates the negative-going portion of signal and squares off the positive-going portion of the signal and inverts it, to produce the signal shown in the graph of Figure 24. That signal appears at a point 257 of Figure 21. It cuLLe ~,~u..ds to the WX signal at terminal 220 of the simplified diagram of Figure 18. Terminal 257 is c~nnPctPd to a transistor 259 in such a way as to perform a logical NAND function. The output signal, at terminal 261, is shown on the graph of Figure 27.
On Figure 20, the reference channel of terminal 241 goes to a b~ndr~ filter 243. one output of the filter 243 goes via a tPrm;n~l 250 to an amplifier 263 as shown on Figure 21. The SPST switch 249 is controlled by the transistor circuit 263 and hence by the reference signal at 253. That reference signal turns on the cross-20 wire signal channel 251 ~hen a strong reference signal is present and positive. (See Figure 24).
The reference-channel b~ndr~c filter 243 also outputs a signal through a diode 265 to an inverting input tPrminAl 267 of an amplifier 269, Figure 21. The 25 waveform at input terminal 267 is shown on the graph of Figure 25. It is a negative-going signal whose magnitude increases as the vehicle approaches the center of the cross wire and whose magnitude dimlnich~e as the vehicle continues past the center. It is the algebraic sum of 30 the outputs of the front and rear coils.
At a threshold of minu6 1. 2 volts the ref erence signal at 267 is tripped. Amplifier 269 is configured as a Schmitt trigger with about 0 . 2 volts of hysteresis .
The threshold for decreasing magnitude is 1. 0 volt, as 35 shown in Figure 25. This threshold is passed as the vehicle continues f orward past the wire cross . The output of the amplifier 269, at a tPrmin~l 271, i8 shown =~ .
=.
t as a large sguare graph 293 in Figure 26.
The sguare graph 293, which has a range from negative 11 volts to positive 11 volt6, is applied - through a diode 273 and a resistor 275 to the base of 5 transistor 259. That signal servefi as the reference-channel input to the NAND gate whose principle ^nt is transistor 259. Transistor 259 is part of the NAND
gate 221 of the simplified diagram of Figure 18.
The circuit of Figure 22 has hAnArA~:s filters 277 ^nd 279, both of which are tuned to 1155 ~z for passive loops. Otherwise, the circuit of Figure 22 is identical to that of Figure 21. The output of circuit 22 is at a point 281. This is the cross-wire signal output when a passive loop is used instead of an active guidewire.
The curves of Figures 23 through 27 are aligned vertically over each other to provide the same vehicle-position scale on the abscissa for ~11 of them.
Collectively they portray what happens in the circuit when a guided vehicle having antennas 205-215 as in Figure 7 enters a t~minAl and drives over a wire-crossing that it must detect for purposes of longit~lA;nAlly positioning the vehicle. The Ah~riRsa of all of the graphs of Figures 23 through 27 is distance expressed in inches, as measured positively and negatively from a zero point 283 on Figure 23. Point 283 is the vehicle's position when the middle coil 207 is directly over the wire-crossing on the floor.
As shown in Figure 23, at a distance of -3 inches, a curve 285, which is the wire-crossing signal at t_rminAl 253 of Figure 21, has increased to a +0.2-volt level. A Schmitt trigger 255 trips its output from positive saturation level to negative saturation level 289 at a point 287 in Figure 24. The graph at Figure 24 is the signal at t^rminAl 257 of Figure 21, as a function of the vehicle ' 5 longitudinal position.
On Figure 23, when the curve 285 decreases (at WO 92~9941 ~ PCr/US91/08892 2~9~2 a short distance to the right of the zero-point 283) to a level more negative than -0 . 2 volts, which is the negative threshold level os Schmitt trigger 255, the output signal at terminal 257 returns to a positive 5 saturation level. The 6ignal 289 e6sentially serves as one input of the NAND gate 259.
Turning now to the reference signal channel of Figure 21, a signal at tPrminAl 267 dim;nichPq gradually from zero to a minimum at the WiL~z _Lossing center 10 ~c:~uLe~Led above by point 283. The waveform at f~rmi 267 of Figure 21 is the V-6haped waveform 291 of Figure 25. A6 the signal 291 decreases past -1.2 volts, amplifier 269 is triggered to 6aturate to the positive rail. Alternatively, as signal 291 increases past -1. 0 volt, amplifier 269 is triggered to saturate to the negative rail. The output signal at tPrminAl 271 is shown a6 waveform 293 in Figure 26.
At the output tPnminAl 261, a negative-transition pulse 295 is produced at a wire cross. As shown on Figure 27, its leading edge 296 occurs at a place very slightly more positive than the zero center point 283 of the wire 87 on the floor. Its positive-going edge, if the vehicle were to continue in a forward motion, would occur at a position 297 on Figure 27. The output signal at tPrminAl 261 is a positioning signal whose edge 296 indicates that the middle coil 207 is almost directly over the WiL~ _Lu~Ling. This signal goes to the motion control ~Lùcessur 61 to stop the vehicle and/or control its repositioning, by means of well-known 3 0 computer control ~ruyL i nq techniques 2!!PA~ of Hea~q i nr~ of Vehicle . in One ~h~rl i ' of the Invention In one preferred Pmho~li L, one sensing antenna 47 is mounted at the front of the vehicle and another sensing antenna 47A is mounted at the back of the vehicle, as shown in Figure 11. A mea~uL ~ t of the lateral offsets of the center of each of the antennas 47 W092/09941 ~ 2~9544a PCI`/US91/08892 nnd 47A from a central longitudinal wire segment 307 on the f loor indicates the vehicle ' s heading . The net difference in offsets divided by the longitudinal spacing . 308 between the antenna assemblies 47, 47A i6 the tangent 5 of the heading angle of the vehicle relative to the wire 307 .
The signals from antennas 47 and 47A are processed in the manner described in detail above and subtracted in a comparator 309 and entered into a portion 61A of the miL;~ul __~er 61. See Figure 11. Stored in the mi~:L. __Ler 61 is information as to the longitudinal spacing 308 between the two antennas, which enables the computer to compute the vehicle ' s heading .
~ltern~tive Receivincr Svstem lq'mKn~ Iavin~
Ph~ Locked LOODS
An alternative ~ of the apparatus f or ~lot~ mi ni n~ lateral position of a vehicle inCu~yu~tes a phase-locked loop (PLL). Figures 28 and 29 show this ~mhor~ which is an AC biasing system for c -- ting for (i.e., subtracting) the ~: L of signal that is received at antenna 91 directly ~rom the transmitting antenna 71.
The filtered and amplified signals from the lateral receiving antenna 91 are at torm;n~l~ 153, 155 of Figures 15 and 28. The left-side signal at 153 is input to PLL 313 and subtracted in a summing amplifier 317 from the output (at 325) o~ the PLL 313, as shown in Figure 28. The difference is a voltage at tormin~l 321, whose ~mplitude is approximately proportional to the vehicle 15 lateral position. The PLL 313 is shown in more detail in Figure 29. The output 325 of the PLL 313 is the output of a sinusoidal voltay- _u.,~,ulled oscillator (VCO) 327, which is part of the PLL 313, as made clear by Figure 29.
The right-side signal at tormin~l 155 is processed by similar circuits.
The VCO 327 ~rudu~ s a signal whose pha6e is . locked to the phase of the input signal 153. This is W0 92/09941 ` ` 2 ~ 9 5 4 ~ 2 - ~ PCriUS91/08892 ~

accomplished by multiplying the output of oscillator 327 (as modified by a gain CUII~LU1 circuit 329, under control of DC voltage at a tPrmin~l 337) in a multiplier 339.
The output of 339 is a DC signal repre5entative of the phase difference, or phase error, between the output of the VC0 327 and the input signal 153, and is attempted to be driven to zero by the PLL.
This DC signal enters a lowpass filter 341, whose output at 342 i5 used to control the phase of the VC0 327, (the 06cillator's frequency being the time rate of change of its phase). This arr;-n~, L provides a final output signal at tPrm;n~l 325, which is a robust AC
signal having the same phase as that of the input signal 153. The circuit of Figure 29 is block 313 of Figure 28.
To make this alternative ~ ` - '; r t more refined, the automatic gain control 329 is employed during initialization to set up the amplitude of the output of the VC0 327 to be egual to the signal voltage at the tPrm;n 1l 153 under conditions described below.
2 0 The operation of the P~horl; r ~ shown in Figures 28 and 29 is as follows. The phase of the left signal is tracked by the PLL 313 of Figure 28 (and the phase of the right signal is tracked by a corrpcpnn~l i n~
PLL). The PLL 313 provides at its output 325 a signal of pre-adjusted amplitude (which is set upon initializa-tion), and of phase that tracks the phase of the received signal at terminal 155. -Inltialization is performed far away from floor wires. The only input signal at that time is that which is induced directly in antenna 91 by a magnetic field produced by the transmitting antenna 71. To initialize the system a switch 331 is closed and a motorized potentiometer 335 (or alternatively an up/down counter and a D/A converter) are adjusted to achieve a DC level at tPnm;n~l 337 such that the output signal of the PLL
313 is at a certain amplitude. That certain amplitude is the value at which the PLL's output signal 325 is exactly -_ _ _ _ _ _ _ _ _ _ _ . . _ _ _ _ _ _ . .

WO 92/0994] PCr/US9l/08892 `2~ ~44a ^--elaual to the input signal 155 as det~rmin~l by the summing amplifier 317.
The switch 331 i6 then opened. The motorized ~- pot 335 remains in the position in which it was set during initialization. It continues to control the gain of block 329 via terminal 337 so that the amplitude of the output signal at 325 from the PLL 313 remains the same as it wa6 at initialization. If the signal 155 changes in amplitude, the lateral position signal at output 321 changes.
The signal at t~rm;nAl 321 can be h;~n~rs~cc filtered, fullwave rectified, subtracted from the signal of the right-side receiving antenna, and used for control in the same manner as is shown starting with t~rm;n~l 155 in the ~mhor~ L of Figure 14.
Alternative Transmit~;n~ ~ntenna Placement Figure 3 0 illustrates an alternative technique for passive loop positioning of a vehicle in a t~rm;nAl that is eSIuipped with a passive loop. The passive loop 343 in this ~mho~l;- L is a coil of wire with its ends connected together 50 as to form a closed loop, and which is flopped over at a point such as point 345 50 that it forms a left-hand loop 347 and a right-hand loop 349.
Magnetic f ields produced by current in the loop reinforce, i.e., they are additive, in the center leg 350 where two wire segments lie close to each other. The transmitting antenna system comprises two antennas (coils) 351 and 353, on separate cores, which are spaced apart by an amount that places them over the outside legs of the folded loop 343. The coils 351 and 353 are phased 80 a6 to reinf orce each other in inducing current in the loop 343. The receiving antenna assembly 355 is the same as was described earlier.
Another F~hoAj--?~t An alternative ~ho~l; L of the tc~rm;n~
positioning mode of the vehicle navigation and guidance apparatus processes the recelved signals dlfferently than _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ = _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ WO92/09941 ` '20 ~ 5 ~ 4 2 described above . This alternative PTnhO~ i L is adequately describable without a separate f igure . It has equipment that ~;ubtracts the two rectif ied signals that come from the rectifiers 113, 115 of Figure 12A. Their 5 difference is a voltage approximately proportional to the lateral position of the vehicle. In this ~ it i5 best for the two direct signal6 from the transmitting antenna assembly 71, which are received by the two receiYing coils 95, 97, to be of equal ~L~ y~.
10 Equality of direct signals is achieved by adjusting the position of the receiving antenna assembly 91 with reEipect to the transmitting antenna assembly 71.
Automatic Bias-Sett i n~7 Embo~ i An alternative P~ho~ provides automatic setting of the biase6 129A and 129B of Figure 12B; such automatic setting is a calibration step for the Proportional Positioning System. Bias setting ~es for an undesired offset of the receiving antenna's signal (see Figure 7), caused by energy that is directly magnetically coupled from the transmitting antenna 71 to the receiving antenna 91 (i.e., energy not received via the passive loop 55).
The Proportional Positioning System is a portion of the AGV de~;cribed elsewhere herein. It inrl~ Pl:, as shown on Flgure 4B, the on-board magnetic transmitter 68, the passive loop 55, the lateral-position antenna of block 47, the lateral-cllannel preamplifier 109, the Passive Lateral SllhrhAnnPl including tPrTn;nAl~
113A and 118, and the A/D converter 135.
The preamplifier 109 is shown again in the more detailed schematic diagram of Figure 12A, whose circuit is continued on Figure 12B. Figures 12A and 12B, whose output is at tPr~n;nAl 118, depict only a manual bias-setting circuit, 129A and 129B.
rn the alternative currently preferred automatic ~ now to be described, the signal at a bias tprminAl 361 is an automatically controlled zero-to-:-.

WO92/09941 ~ 12 PCI/US91/08892 five-volt bias for offsetting the direct magnetic coupling c _ ,rnt from the transmitting antenna 71. The nutomatic bias-setting circuit as a whole is a closed ~- loop that, during calibration, provides whatever voltage is nPcP~ ry at tprmin~l 361 to make the voltage at tPrm;nAl 362 equal zero.
The automatic bias-6etting circuit makes precise adjual -nt of the location of the transmitting antenna ~ i.PC'P~ y and enables easy _ ^n~tion for aging of ~ -nts, etc. Circuitry of this type is preferably provided for both of the receiving coils 95, 97 .
The . _ -nts of the automatic circuit and their into:, ~;om.e- Lions are shown in the circuit diagram of Figure 31. To show how the circuit interfaces with the other AGV circuits, the top line of Figure 12B is reproduced as the top line of Figure 31, except with the automatic bias circuit replacing the manual ~ias circuit 12 9A .
As s~en in Figure 31, an analog signal at output terminal 362 of amplifier 117 is conducted to a digitizing circuit 366, which consists of an inverting amplifier and a transistor clipping circuit. Circuit 366 produces a logic 1 level at its output terminal 363 if the signal at terminal 362 is positive, and a logic 0 level at tPrmin~l 363 if torm;n~l 362 is negative.
Terminal 363 is connected to a counter 372, which also has a clock input tPrmin~l 364 for receiving pulses that are to be counted. The direction of counting is detPrminPd by the logic level of tPrm;n~l 363. The count is incremented upon OCI.:UL L c:~.ce of a clock pulse if tPrm;n~l 363 currently has a logic 1, and decremented if tPrm;n:ll 363 has a logic 0. Counter 372 is a model 74A5867, manufactured commercially by the Company of Texas Ina-~, -nts Inc., Dallas, Texas, 75265.
Another subcircuit 368 performs the function o~
generating clock pulse signals at a controllable _ _ _ _ _ _ _ _ _ _ _ _ _: _ _ _ _ _ _ _ _ _ _ _ . .

WO 92/09941 ` ~, O 9 !~i 4 ~ 2 PCT/US91/08892 ~

frequency. The absolute-value circuit 368, whose input i8 at terminal 362, provides an analog voltage at an output terminal 369. The analog voltage at 369 is the magnitude of the signal of terminal 362, so t~r~;nAl 369 5 is never negative, irrespective of the polarity of the bipolar signal at torminAl 362.
TP~n~;nAl 369 is connected to a voltage-controlled digital oscillator 370; it produces output pulses at a frequency that depends upon the control voltage at terminal 369. The oscillator 370 provides output pulses at a terminal 364, which are cnnrl~ll~ted to the clock input t~rm;nAl of the counter 372. The oscillator 370 is a model NE555, manufactured ~;lally by Texas InDLl, Ls, Inc., Dallas, Texas, 75265.
The count contents of the counter 372 are connected to an EEPROM (Electronically Erasable PLoyr hle Read-Only Memory) 374, which is optional in this circuit. The EEPROM i5 capable of storing the count when it is - nr~r~d to do so by the outer loop mi~:~ u~JLocesso~ 67 . The output of the EEPROM is connected to a (digital-to-analog) converter 376, which is a model DAC0808, manufactured commercially by National Semiconductor Company of Santa Clara, California, 58090.
The analog output of the D/A converter 376 is inverted in an amplifier 378, whose output is connected to the bias terminal 361 of the amplifier 117.
Operation of the circuit is as follows. The calibration process is performed at a time when the Yehicle is not over a wire. At such a time antenna 91 i5 not receiving any ~ - L of signal via wires on the ground. To start a calibration (bias setting) the outer loop mi.:~u~l~,cessor 67 sends a calibrati~,-. nr~ bit to an "enable" terminal 365 of the counter 372.
If the voltage at t~nm;nAl 362 is negative, the binary signal at t~r~;nAl 363 is low, which causes the direction of counting of the counter 372 to be downward.

W092/09941 2 0 ~ ~ 4 4 ~ PCr/US91/08892 ; ~ -The decrea6ing count passes through the EEPROM 374 and causes the D/A converter 376 to receive less input current, causing the voltage at the bias terminal 361 to increase. That makes the voltage at tPrminAl 362 less 5 negative, so the 362 voltage moves toward a null.
Conversely, if t~rminAl 362 is po6itive, the signal at tQrminAl 363 goes high, which causes the counter 372 to count upward, and causes the D/A converter 376 to receive more input current, causing the voltage of 10 t~rminAl 361 to decrease. Thereupon, the voltage at inAl 362 decreases toward zero.
The frequency of pulses at the clock input t~rminAl of the counter 372 depends inversely upon the magnitude of the voltage at t~rm;nAl 362; a greater 15 magnitude results in a greater frequency of the pulses that are counted by the counter 372. Consequently the offset calibration 6ignal at t~rm;nAl 361 approaches a final value faster when it has farther to go. It reaches a final value when the voltage at torm;nAl 362 is zero, 20 which reduces the counting rate at t~m;nAl 364 to zero.
The counter 372 retains its count contents, ~o the proper bias voltage remains on the bias t~rm;nAl 361.
If the optional EEPROM 374 is provided, the vehicle need not be calibrated anew every time it is 25 started. After a calibration the EEPROM i6, ntl~d by the outer loop mi-Lv~Luce6sor 67 to read the output of the counter 372 and store the value in its memory. The EEPROM theref ore can reproduce the count that was in the counter 372 just before the power was turned off, and if 30 it is still an appropriate value the calibration need not be repeated.
The following table comprises a list of - -nts and eat types or values for circuits seen in Figures 6, 12A, 12B, 15, 16A, 16B, 17A, 17B, 18, 19, 20, 21, 22, and 31:

W092/0994~ ~ 20 ~944a PCr/US9l/08892 Name Value or Tvoe C3 Capacitor . 22 ~f C6 Capacitor . 22 I~f Cg r:~r~citor .1 ~Lf 5 Cg ' Capacitor . ol ~Lf Clo Capacitor 2 . 2 l~f C12 Capacitor 1 ,uf C15 Capacitor . 27 ~f C16 Capacitor . 22 ,uf 10 C17 ' Capacitor 1 ~f Cls Capacitor lo ~Lf C20 Capacitor 10 /~f C21 Capacitor lo ~Lf C22 Capacitor lo ~f 15 C23 Capacitor 10 ~f C24 Capacitor lo ,uf C25 Capacitor . 847 ~Lf C27 Capacitor lo ,uf C28 Capacitor . 047 ~f 20 C28 ' Capacitor 8 . 8 l~f C29 Capacitor lo ~f C31 Capacitor .1 ~f C31 ' Capacitor 8 . 8 ,uf C32 Capacitor 10 ,~f 25 C34 Capacitor .1 ,uf C35 Capacitor .1 ~lf C38 Capacitor lo ,I~f C38 ' Capacitor 2 . 2 ,uf C4 1 Capacitor . 2 2 ,~f 30 C47 Capacitor 2.2 ~f C48 Capacitor 4700 I~f C49 Capacitor . 022 ILf C50 Capacitor .1 ~f C51 Capacitor . 0047 ILf C52 Capacitor . 0047 ,uf C59 Capacitor .847 ~f C60 Capacitor ~47 WO 92/099A1 ~ 4 2 PCT/US9l/08892 C61 Capacitor . 22 ~f C67 Capacitor 2 . 2 ~Lf C68 Capacitor 4700 ~Lf . C69 Capacitor . 022 llf 5 C70 Capacitor .1 ~lf C72 Capacitor 10 ILf CRl Diode lN4148 CR2 Diode lN4848 CR3 Diode lN5234 10 CR4 Diode lN5234 CR5 Diode lN4148 CR6 Diode lN4148 CR8 Diode lN4148 CR10 Diode lN4148 CRll Diode lN5234 CR12 Diode lN5234 CR13 Diode lN4148 CR14 Diode lN4148 CR17 Diode lN4148 CR18 Diode lN4148 CRl9 Diode lN4148 CR23 Diode lN4148 CR24 Diode lN4148 CR25 Diode lN4148 2 5 CR2 6 D iode lN4 14 8 CR27 Diode lN4148 CR28 Diode lN4148 CR29 Diode lN4148 CR3 0 D iode lN4 14 8 3 0 CR3 1 D iode lN4 14 8 CR32 ~ Diode lN4148 Ul Switches/Gates LF11202D
U8 Switches/Gates LF11202D
U12 Switches/Gates 7402 U13 _ Switches/Gates 7402 , WO 92/09941 ~ 4 4 2 ~ ~ PCr/US91/08892 U14 Switches/Gates LF11202D
U17 Switches/Gates LF1202D
U24 Switches/Gates LF11202D
U29 Switches/Gates LF11202D
U30 Switchefi/Gates 7404 Ll Tnr'll~tors 50 m l L2 Inductors 50 m~
L3 Inductors 50 m3 L4 Inductors ~ 72 . 4 m.~
10 L7 Inductors 72 . 4 m,l L8 Inductors 50 m.
Ls Tnr'17~t~rs 50 m.
L10 Tn~ 7rs 50 m.~
Lll Inductors 72 . 4 m l 15 L12 Inductors 72 . 4 m~
El Jumpers E2 Jumpers U2 Oper . Amp . LF3 4 7 U4 Oper . Amp . LF3 4 7 2 o U5 Oper . Amp . LF3 4 7 U5 ' Oper . Amp . L'.~675T
U6 Oper . Amp . L'q675T
Ull Oper.Amp. LF347 Ul9 Oper.Amp. LF347 2 5 '~2 0 Oper . Amp . LF3 4 7 'J2 3 Oper . Amp . LF3 4 7 U2 5 Oper . Amp . LF3 4 7 U2 6 Oper . Amp . LF3 4 7 U2 8 oper . Amp . LF3 47 30 R2 Resistor 301 Ohms R3 Resistor 3 . 57K "
R4 Resistor 63 . 4K "
R4 ' Resistor 165K "
.. ~

WO 92/09941 2 ~ 9 ~ ~ 4 ~ PCI`/US91/08892 R5 Resistor lK "
R5 ' Resistor 15K
R6 Resistor 1. 4K
- R6 ' Resistor lK
5 R7 Resistor lK
R7 ' Refiistor 301 R8 Resistor lK
R8 ' Resistor 3 . 57K
R9 Resistor 12 . lK "
R9 ' Resistor 80 . 6K
R10 Resistor lK
R10 ' Resistor 165K
Rll Resistor 165K
Rll ' Resistor lOOK
R12 Resistor lOK
R13 Resistor } . 4K
R13 ' Resistor lOK
R14 Resistor lK "
R14 ' Resistor lOK "
R15 Resistor lK
R15 ' Resistor lOOK
R16 Resistor 27.4K
R16 ' Resistor lOK
R17 Resistor 165K
R18 Resistor lK
R18 ' Resistor 69 . 8K
R18 " Resistor lOK
Rl9 Resistor lOK
Rl9 ' Resistor 499 R20 Resistor lK
R20' Resistor 33.2K
R21 Resistor 38 . 3K
R21 ' Resistor lM
R22 Resistor 80 . 6K
R22 ' Resistor 165K
R23 Resistor lK
R23 ' Resistor lOOK

- .
WO 92tO9941 PCr/US91/08892 2095~42 R24 Resistor lOK
R24 ' Resistor 100 R25 Resistor lOK
R25' ' Resistor 100 5 R26 Resistor lOK
R26' Resistor 22.1K
R27 Resistor lOK
R27 ' Resistor 121 R28 Resistor lOK
10 R28 ' Resistor 100 R29 Besistor lOK
R29 ' Resistor 200K
R3 0 Resistor lOK
R30 ' Resistor llOK
15 R31 Resistor 20K
R31' Resistor 51.1R
R32 Resistor 2 . 74K
R32 ' Resistor 4 . 99K
R33 Resistor 1 1/2 watt "
20 R33 ' Resistor 8 . 06K
R34 Resistor R34 ' Resistor lK
R35 Resistor 9 . lK
R36 Resistor 1. lK
25 R37 Resistor llOK
R38 Resistor lOK
R39 Resistor lOK
R40 Resistor R40' Resistor 221K
30 R41 Resistor 127K

R41' Resistor lOK
R42 Resistor 15. 4K
R42 ' Resistor lOK
R43 Resistor lOK
35 R44 Resistor lK
R45 Resistor lOK
R47 Resistor 604K

WO 92/09941 2 ~ 9 5 ~ 4 ~ PCr/Usg1/08892 R49 Resistor lX
R50 Resistor lOK "
R51 Resi6tor 49 . 9K "
~- R52 Resistor lOK "
5 R53 Resistor lOIC
R54 Resistor 604K
R55 Resistor 49 . 9K "
R56 Resistor 20K
R57 Resistor lK
10 R58 Resistor 20K
R59 Resistor lK
R60 : Resistor lO0 "
R61 - Resistor lK
R61 ' Resistor lR
15 R62 Resistor 137K
R62 ' Resistor 100 R63 Resistor 15K
R63 ' Resistor 100 ~' R64 Resistor 1. 4K "
2 0 R65 Resistor lK
R66 Resist~r lK "
R67 Resistor 12 . lK
R68 Resistor 137K "
R69 Resistor 137K
25 R70 Resistor lOK
R71 Resistor 1. 4K "
R72 Resistor lK "
R73 Resistor lK
R74 Resistor 27 . 4K "
30 R75 Resistor 137K "
R76 Resistor lOK "
R77 Resi6tor lOR "
R78 Resistor lK
R79 Resistor lM
35 R80 ' Resistor 165K "
R81 Resistor lOOK "
R81 ' Resistor ~l00 "
-~ , --W092/09941 ~ ~ PCr/US91iOX892 2~95442 R82 Resistor lOK
R83 Resistor lOK
R84 Re6istor lOK "
R85 Resistor lOK
5 R86 Resistor lOK
R87 Resistor 200K
R88 Resistor lOK "
R89 Resistor 20K "
R90 Resistor 4 . 99K
10 R91 Resistor 8. 06K
R92 Resistor lK
R97 Resistor lOOK
R98 Resistor lOOK "
R107 Resistor 845K
15 R108 Resistor 165K "
R109 Resistor 1. 4K
RllO Resistor 15K
Rlll Resistor 13 . 3K
R112 Resistor 9 . 9K
20 R113 Resistor 165K
R114 Resistor 845K
R115 Resistor 4 . 99K
R116 Resistor lOK "
R117 Resistor lOK "
25 R118 Resistor 9. O9K
Rll9 Resistor 4 . 53K "
R120 Resistor 27 . 4K "
R121 Resistor 56 . 2K "
R122 Resistor 22 . lK "
30 R124 Resistor 28.7K

R126 Resistor 28.7K "
R127 Resistor 8 . 06K "
R128 Resistor 25. 5K
R129 Resistor 23.2K "
35 R133 Resistor 25.5K "
R134 Resistor 8 . 06K "
R13 6 Resistor lOOK "

W092/09941 ~ ~0~!~44a'! PCr/US91/08892 R137 Resistor 49 . 9K "
R118 Resistor 49 . 9K "
R13 9 Resistor 10 OK
R141 ~ Resistor lOK "
5R142 Resistor lOK "
s R143 Resistor lOK "
R144 Resistor lOK "
R145 Resistor 100 "
R146 Resistor loo 10 R147 Resistor lOK "
R148 Resistor lOK "
R149 Resistor lR "
R150 Resistor lOK "
R151 : Resistor lOK "
15 R152 Resistor 100 "
R153 ' Resistor lR "
R154 Resistor 35.7K
R155 Resistor 35.7K "
R156 Resistor 1. 21K "
20 R157 Resistor 15K "
R158 Resistor 15K "
R159 Resistor 47 . 5K "
R16 0 Resistor 82 . 5K "
R161 Resistor &45K "
25 R162 Resistor 165K "
R163 _ Resistor 1. 4K "
R1~7 Resistor 165K "
R168 Resistor 845K "
R164 Resistor 15K "
30 R165 Resistor 13 . 3K "
R166 Resistor 49 . 9K "
R169 Resistor 4 . 99K "
R170 Resistor lOK
R171 Resistor lOK "
35 R172 Resistor 9 . O9K "
R173 Resistor 4 . 53K "
R174 Resistor lOK "
_ WO 92~0994l PCr/US91/08892 2~9~4-42 R175 Resistor lOK "
R176 Resistor 27 . 4K "
R177 Resistor 56 . 2K "
R178 Resistor 22. lK
5 R180 Resistor 47 . 5K
R181 Resistor 47 . 5K "
R182 Resistor 3 . 32K "
R183 Resistor lOK
R183 ' Resistor 27 . 4K
10 R184 Resistor lOOK "
R185 Resistor lOOK "
R187 Resistor lOK "
R189 Resistor lOK "
Rl90 Resistor 150K "
15 Rl91 Resistor lOK "
R192 Resistor lK "
R200 Resistor lOK "
R201 Resistor lOK
R202 Resistor lK "
2 o R2 0 3 Res i ~tor 2 OK "
R204 Resistor lK "
R205 Resistor 100 "
R206 Resistor 2K
R207 Resistor 2K
25 R208 Resistor 2K "
R209 Resistor lOK "
R210 Resistor lOK "
R211 Resistor lOK "
R212 Resistor lOK "
30 R213 Resistor lOOK "
R214 Resistor lOK "

R215 Resistor lOK
R216 Resistor lK "
R217 Resistor 3 . 3K "
3 5 R218 Resistor lOK "
R219 Resistor 3 . 3KR ~I
, i; -'`'` ~('r-/lJS91/08892~

Ql Transistor 2N2222 Q2 Trans istor 2N2 2 2 2 Q3 Transistor 2N2222 ~- Q4 Transistor 2N2222 Q5 Transistor 2N2222 ..

WO 92/09941 ~ PCr/US91/08892 2~95442 102 VPhinle Naviaa~ion i~n~l GuidancP
A6 earlier A1~SclosPA~ vehicle 2A comprises a plurality of navigation and g~1iA~nne systems. Under control of AGVC 13, vehicle 2A selectively guides over 5 guidewire routes 3, in tPrminAl~ ll, and along a ground marked route 5, performing aut~n or self-contained guidance between ground markers 6. G~ nne along a guidewire is known in the art and will not be further covered herein.
Aut~ or Self-Contained G~ nnP
In the currently preferred: ' :'i L, each vehicle 2A uses feedback from a linear encoder 58 from each wheel f or aut~nl ~ or self -contained guidance .
Aperiodic mea,,uL~ L of position and direction from 15 update markers 6 and the angular rate sensor system (commonly called gyro 500) provide sufficient rPtll~n(1;1nry of ~ L to correct positional and directional errors and allow the allocation and application of real time calibrations to correct for angular rate sensor 20 dri~t, t~ U, t changes, aging and wear of linear mea:,u- L Ls, and the like. In the currently preferred Pmho~ L~ the inertial g~ nne sy6tem provides a vehicle ~11irl~n-e accuracy of an error having a standard deviation of 2 inches over travel of fifty feet 25 between ground markers 6. While such accuracy is not sufficiently accurate for travel within a ~Prmin 1l, it is adeauate for travel on the floor of a facility. As described earlier, tPrminP~1 positioning guidance of the currently preferred: ` ir- L provides a maximum error 30 of + l/4 inch.
A block diagram of navigation and q--if~nne system 800 is seen in Figure 56. Note that the contents of Figure 56 comprise all of the elements of Figure 4A
plus an outer loop comprising an update marker system 35 (UMS) block 400, an angular rate 6ensor block (gyro block 500), and an outer loop ~)LO~ Ss~JL block 67. The outer loop ~ ay be considered to compri6e the navigational _ WO 92/09941 ~ 2Q ~ 5 i~ Z PC~uS9~08892 system while the inner loop (that which resides inside the outer loop) may be cnnci~red to comprise the guidance system.
Thus, the -rts of the outer loop are used 5 to aperio~; c~l ly provide mea~u.l L of position and ~1rer~1On as vehicle 2A travels across a marker 6 and reads direction f rom gyro 500 to provide updates from time to time. After readings are taken from blocks 400 and 500, a Kalman 10 filter calculation is made whereby the guidance position and direction are updated and real time calibrations are made. Kalman calculations and calibrations are described in detail hereafter.
l'h~ Tn~rtial Platform The inertial platform providefi a source of angular ~ s which are used in combination with estimates of vehicle 2A position from a ground marker 6 to update the AGV 2A control system. In combination, the ground marker and angular mea- u.~ Ls need to provide 20 sufficient precision and accuracy to maintain an acceptable guidepath error between each update. In the currently preferred ~mho~l1r -~, the acceptable guidepath error has a standard deviation of 2 inches in fifty feet of travel between ground markers 6. It is of primary 25 ; Lelllce that the combination of angular updates, inputs from the wheel encoders 58, and ground marker 6 position detcrminations, made Eurc~cively, provide sufficient L~ Anry of vehicle 2A position and heading information that the errors due to deflrl~nri~c of the 30 measuring devices comprising change~ due to aging, wear, drift, and t~ ~turla, are c uLLc ~:l.able by real time calibration using Kalman filtering. Reference is made to Figure 57 wherein the major elements of an inertial platform (commonly referred to as gyro 500) are seen.
35 The major elements of gyro 500 comprise a printed circuit board 904 which contains gyro 500 control loop circuits, the angular rate sensor 900, the inertial table 700 -- .

WO 92/09941 ~ 2 0 9 ~ 4 4 2 PCr/US91/08892 comprising a motor 916 which con~inllnllcly drives the angular displa~ - ~ of angular rate sensor 900 to a null position, an angular rate to electrical signal encoder 58, and a slip ring assembly 906 and are centrally 5 mounted in well 26 of vehicle 2A.
A package 901 (see Figure 61) comprising heaters and insulators completely ~nl -F~ angular rate sensor 900 and is affixed to support 992 which is shock mounted to inertial table 700 with stand-offs 912. On 10 the opposite side of inertial table 700, printed circuit board 904 is firmly affixed in vertical orientation. A
shaft is centrally fl i RpssP~l through and connected to moving parts of the gyro 500 comprising slip ring assembly 906, a hub 994 which firmly supports inertial table 700, a motor rotor 922 (seen in Figure 60), and the moving parts of encoder 88 ' . Wires and other parts, such as circuit , L details and power supply parts are not shown for clarity of presentation.
Electrical signals are trans~erred from the moving parts of inertial table 700 to n~ ving parts through slip ring assemoly 906. Of the five slip ring connections seen in slip ring assembly 906, four are used in the currently pref erred P"~ho~ L . Non-moving parts of slip ring assembly 906 are supported by a nylon bracket 910 attached to an upper housing member 926, only partially seen in Figure 57. Support for the inertial table is provided by mounting bracket 924, better seen in Figure 60.
A block diagram of the gyro 500 is seen in Figure 63. A signal comprising the rate of angular change is sent to a network of amplification and -, Lion circuits 998 wherefrsm feedback current to drive motor 916 is provided. Motor 916 is driven to maintain angular rate sensor 900 in a null direction.
The angular travel of motor 916 is sensed by encoder 88 ' wherefrom a signal is provided to outerloop p~ ~ cessvI 67 for Kalman filtering and other proCPRRin~.

, . !
:'.

2()g~ll42 PCr/US9~/08892 Selection and prorPctsin~ the output of an angular rate sensor f or nn AGV vehicle is not trivial .
All angular rate sensors drift or diverge as a function of time. A6 an example, navigation angular sensor drift rates are commonly in the range of . 01 degrees/hour, t submarine angular sensor drift rates are commonly more restrictive, in the range of . 001 degrees/hour, while angular sensors may have as high a drift rate as 100 to 1, 000 degrees/hour. The cost of an angular rate sensor normally increases significantly with decreases in rate of drift. The cost for very low drift rate angular rate sensors can be as much a6 the cost of an entire AGV 2A.
In addition, the angular rate sensor for an AGV
2A must have a rapid warm-up or response time. Some angular rate sensors, such as gas gyros, require up to one-half hour for warm-up. Maintenance of conventional angular sensors is also a concern. The common mean-time-bt:L~ . E_L vicing is commonly under 5000 hours for conventional angular 6ensors.
From a cost and maintenarce ~L~.~ecLive, the angular rate sensor ~elDrt~ for the currently pre~erred '~ L i8 suitable to the requirements of the invention. While other angular rate sensors can be used in the invention, the selected sensor is from a family of rate sensors (ARS-C121, ARS-C131, and ARS-C141) provided by Watson Industries, Inc., 3041 Nelby Road, Eau Claire, Wisconsin, 54701. Each of the family of rate sensors mentioned above provide full scale outputs at 30, 100, and 300 degrees/second, respectively. Nodel ARS-C121 is the sDl ~rted product for the currently preferred nt of angular rate sensor 900 because it provides the greatest sensitivity over the range required. Use of the selected rate sensor requires the ::VllCULLttnL
implementation of the inertial table 700 to eliminate the possibility of saturating measuring Ls.
The selected angular rate sensor 900 is an entirely solid state, "tuning fork", single axis sensor ,, - ~
_ _ _ _ _ _ _ WO 92/09941 2 0 9 ~ ~ 4 2 PCr/US91/08892 nnd utilizes pi~Q~ ic vibrati~g beam technology to produce an inertial sensor with no moving parts. It provides an analog output voltage which is proportional to the angular rate about its sensing axis. At zero 5 angular rate, the output i8 zero volts. Full scale angular rates produce an output of +10 or -10 volts, rl~r~n-l~nt upon direction of rotation. A dual power E;upply, providing regulated +15 and -15 volts, is required .
The Bl'l ecto~ angular rate sensor 900 has a drift rate which, if left ul.~ur~ Led, would make sensor 900 tlmlC5thl~ in the AGV 2A application. Surprisingly, however, use of ~cdu-~-la-~L mea,,UL- - ~ and prQr~csin~
u6ing a Kalman filter to perio~l;rtlly correct for and 15 recalibrate the drift rate of selected angular rate sen60r 900 provides a low cost, ef Eective angular rate sensor for the AGV 2A application.
A feedback control loop - '^l i ng operation of gyro 500 is seen in Figure 58. Physical angular -- v~ t.
of AGV 2A provides positive input ~4, to summing block 968.
Output from summing block 968 is error signal ~c which provides input to function 970. Function 970, G~(s), provides a transfer function approximated by K,/ tl+sr~), wherein K~ i5 equal to a gain of approximately 19 volts/radian/second in the currently preferred The term ( l+sr~) provides a low pasC f ilter with a break frequency of l/r~ equal to 300 radians/ second .
Output of function 970 is a voltage signal, Vs providing input to function 972. Function 972, H(s), converts Vs to a current for input to summer 974. Output from summer 974 is function 978, Kd, which models driving amplifiers for pancake motor 916. The output of function 978 is provided to summer 980 wherein the physical properties of motor 916 are summed with the driving output of function 978. The drive properties of motor 916 are modeled by function 982 as l/R+sL (R and L being WO 92/09941 2 ~ 4 a PCr/VS91J08892 the resi6tive and inductive properties of motor 916).
The output of function 982, Le~l senting motor current, is fedback to function 976 which provides a gain control based upon sensed motor current to 5ummer 974, whereby a better model for control of inertial guidance loop poles s and zeroes i8 provided. The output of function 982 is further provided as input to function 984, which L~L.:sents motor torque. Output o~ function 984 is directed to summing junction 985 which also receives a negative input Td representing torgue disturbances such as stiction. Output of summer 985 i5 cnnnpctprl to function 986 wherein the inertial ~ P~ of motor 916 and inertial table 700 are modeled providing an output representing angular velocity of motor 916. In the motor 916 selected for use in the curren~ly preferred pmhorl;r L, motor inertia is negligible.
The output of function 986 is fedback through gain Ke to summer 980. In addition, output of function 986 (~r) feeds back to summer 968, providing angular table error rate ~de. Further, output of function 986 is detected by an encoder 88 ' (see Figure 63) and fed to a direction and integration circuit 90 which provides input to outerloop E,L~c~ssc,r 67, as seen in Figure 56.
The circuits providing angular sensing and feedback control for inertial st~hi l i 7~tion loop 996 are seen in Figure 59. The output of angular rate sensor 900 is provided to amplifier filter 1012. From amplifier filter 1012, the ~ -- ted signal is resistively coupled to two serially c~nnPctpcl differential amplifiers 1014 and 1016 which provide serially integrated filtering of the signal from amplifier filter 1012. The output of amplifier 1016, is resistively coupled to an inverting amplifier 1018 which comprises variable resistor R28G in the currently preferred ~ n t by which a control of voltage gain is provided. The output of inverting amplifier 1018 is resistively coupled to the inverting input of differential amplifier 1020 and therefrom ~. _ ~
.~ =~=
_ _ _ _ _ _ _ _ _ _ . _ .

W09~/0994~ a~44a~l~ PCr/US91/08892 resistively coupled to the inverting input of di~ferentiaL amplifier 1022 whereby differential drive is provided for motor 916 on lines 1024 and 1026. Three switches RlG prevent output from amplifier 1022, and 5 ahort integrating capacitors C23G and C24G only when the circuit supply voltage is not available, thereby providing an additional degree of control and a delay after the analog fifteen volt supply is available, thereby providing delay which pLC:~GllL~ instabilities 10 which would occur when the motor is driven bef ore the control circuits are operating in a normal fashion.
Types and values of L5 seen in Figure 59 are provided in the following table:
~ 1~ Value or Tv~e R18K Resistor 2M
R19G Resistor 3 .18K
R22G Resistor lOK
R23G Resistor lOK
R25G Resistor 200K
R26G Resistor lOOOK
R27G Resistor 227K
R28G Resistor 6 . 81K
R34G Resistor 227K
R35G Resistor lOOK
R36G Resistor 150K
R37G Resistor R43G Resistor 227K
R44G Resistor lOOK
R45G Resistor SlK
R48G Resistor lOOK
R49G Resistor 150K
R50G Resistor 150K
R51G Resistor 51K
R52G Resistor 150K
R53G Resistor 150K
R54G Resistor 1 1/2W
R55G Resi~toF 1 1/2W

WO 92/09941 2 ~ 9 5 4 4 2 Pcr/us9l/ogg92 R67G Resistor 47R
R66G Resistor 16K
R68G Resistor R69G Resistor 51K
5 R70G Resistor 150K
ClOG Capacitor . 47~LF
CllG Capacitor . 22,uF
C12G Capacitor . Ol~F
C13G Capacitor 15~1F
10 C14G Capacitor lO~LF
C23G Capacitor l,uF
C24G Capacitor lO,uF
C25G Capacitor . 22~LF
C26G Capacitor lOO~F
15 1012 Diff.Amp. LF347 1014 Diff.Amp. LF347 1016 Di~f.Amp. LF347 1018 Diff.Amp. LF347 1020 Diff.Amp. ULN-3751ZV
20 1022 Diff.Amp. ULN-3751ZV
D3G Diode lN4148 D4G Diode lN4148 D7G Diode lN4 0 0 4 D8G Diode lN4004 25 DlOG Diode lN5243B
DllG Diode lN414B
Q4G Diode MCR100-6 KlG Relay Switch DSlE-S-DC12V
A more detailed view of gyro 500 i8 seen in Figure 60. As seen therein, gyro 500 comprises a housing comprising an upper part 926 and a lower part 928. When _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ .

WO92/09941 . ~ ;2 0 ~ 5 4 4 2 PCI/US91/08892 finally assembled the upper part 926 and lower part 928 are releasibly attached together by nut and bolt Al~ iPC 898. Mounting bracket 924 is firmly attached to lower housing part 928. To the upper side 896 of mounting plate 924 a bearing housing 936 is firmly affixed by screws 894, or the like. Attached to the inner bottom side of bearing housing 936 a washer-shaped bearing retainer ring 920 is firmly attached by other screws 894. A bearing 934 tightly constrained between bearing retainer 920 and a snap ring 918 is placed at the bottom of the inverted well provided by bearing housing 936. At the top of the inverted well of bearing housing 936 a second bearing 902 is held in vertical position by a spacer 932. Shaft 908 is held in strict vertical ~ by bearings 934 and 902. Spacer 932 separates bearings 934 and 902 by sufficient distance that any shaft 908 wobble due to any freedom of ~. L in the bearing is negligible.
Shaft 908 centrally CnnnPct~ the inner portion of slip ring assembly 906 to the moving parts of gyro 500 and therewith affixed by a capping nut and washer 890.
Immediately below slip ring assembly 906, shaft 908 is affiYed to hub 994 which comprises an outwardly projecting hub platform support 914, upon which inertial table 700 is ~ecurely affixed. As well, shaft 908 is connected to a motor rotor 922 by a locknut 892.
Finally, the bottom of shaft 908 iB connected to an angular rate ~lPco~l;n~ trAn~ Pr 88'.
As mentioned earlier, printed circuit board 904 is mounted vertically on inertial table 700. Angular rate sensor 900 support 992 is affixed to inertial table 700 by standoff ~Pmhl ;es 912 which interface with support 992 through shock- absorbing gromets 913. Rate sensor 900 is firmly affixed to support 992.
Interconnecting wires 888 are only seen in part extending from angular rate sensor 900 and slip ring assembly 906.
A large mass of wires has been removed from Figure 60 for :
.. ~

WO 92/0994l 2 0 9 5 ~ 4 2 Pcr/us9l/08892 clarity of ~L_ee.-L.tion. Motor 916 is a pancake motor ~lrmly mounted on the bottom side 386 of i n~ bracket 924. The housing for motor 916 comprises an arcuately . shaped concave bottom section 884 and an open centered, washer shaped top section 930. The top section 930 is a t flux return plate which resides above motor 916 rotor 922. Motor 916 is a 12 FP kit motor, part number 00-01281-001, acquired from PMI Motors, Division of R~ l 1 y~ln Corporation, 5 Aerial Way, Syoset, New York, New York 11791.
The central portion of bottom section 884 is modified such that an end bell on motor 916 provides an encoder 88 ' housing mounting connection.
Each spring loaded finger 882 of slip ring assembly 906 is affixed to a mountiLng plate 880 such that the end of each f inger 882 comprises a spring bias causing each f inger to ride c~nn~c~ i vely and continuously in a contact containing groove 879 of the moving portion of slip ring assembly 906. The other end of each finger i8 firmly affixed to mounting plate 880 which is rel~cihly affixed to the upper housing member 926 by mounting plate 910.
For accurate operation of angular rate sensor 900, controlled temperature must be provided. The controlled temperature range for selected sensor 900 is between 55 and 65 Centigrade. As mentioned earlier, it is also important that warm-up time be short. To accomplish a short warm-up time, a novel two heater combination is used. As seen in Figure 61, a multilayer heating/insulating blanket (not shown in prior figures) ~SUL ' .,u.-ds angular rate sensor 900. The multilayer cover comprises, from inside out, a first heater unit 940, a vinyl foam insulating layer 942, a metal foil insulating layer 876, and a second heating layer 944. The insulating layer 942 is adhesively applied to the first heater unit 940 with insulating tape. A Kapton tape, by Dupont is used in the currently preferred ~mho~ L.
_-- .
_ _ . . . _ _ _ _ _ _ _ _ _ _ _ _ WO 92/09941 ~ ~ PCr/US91/08892 ` 112 Metal foil compri6ing a PSA face, placed inward, available from IEPD, Saint Paul, Minnesota is used in the currently preferred: ' 'i L for metal foil insulating layer 876.
First heating unit 940 is controlled by a aLuLe control circuit which is well known in the art by feedback from a t~ aLuLe sensor 948, located internal to heating unit 940 as seen in Figure 61.
Heating unit 944 iB controlled by a similar t~ aLuLe control circuit by feedback from a t~ ULe sensor 946 located internal to heating unit 944. The internal heating unit 940 is d~iqn~d to have a fast heating ui._e time to bring angular rate sensor 900 to t~ aLuLe quickly. The second heating element 944 comprises a greater thermal inertia and is provided to maintain angular rate sensor 900 at t~ c~LuLe over the entire operating period after initial heating by first heating unit 940.
Reference i8 now made to Figure 62 wherein the temperature control curves for the circuits of sensor 946 and 948 are seen. A first curve 952 shows a t~, aLuLe response upon turn-on of sensor 948, showing a rapid rise to a cut-off threshold 960 at time 954 at which t~ aLuLe heater unit 940 is turned off and t~ aLuLè
curve 952 decays toward a cooler t~ aLuLe threshold 964 at which heater unit 940 would be turned back on.
However, as seen by following t~ aLuLe curve 958, before the t~ aLuLe --- èd by sensor 948 falls to threshold 964, second heater unit 944 drives the t- aLuLè of sensor 948 upward above threshold 964 whereupon curve begins to follow the t~ aLuLe path of curve 958 at ULU86UVèl time 956. Second heater 944 control circuits are set to control turn-on and turn-off of second heater at thresholds 961 and 962, respectively.
Because the t~ aLuLe at sensor 948 is not allowed to fall below threshold 962 or to threshold 964, first heater remains in a nu.. u~eLative state after time 952.

WO 92/09941 2 0 9 5 ~ 4 2 Pcr/US91,08892 The Fifth and siX~h Whr~Plc ~S7. 59) Ftnr~ Enr~nr~r 58 As seen in Figure ~ 3, AGV 2A comprises a port fifth wheel 57 and starboard sixth wheel 59. An encoder 58 for each wheel 57, 59 is attached to vehicle 2A in a 5 position to measure individual travel of wheels.
However, in the currently preferred ~ ;r-nt, encoder 58 distances are different than fifth and sixth wheel distances f rom the center 8 6 of the AGV 2A and any given center of a turn 82 and must be cr~nclflr~red in vehicle 10 control calculations by motion control processor 61. ~he following discloses such c~onciflPrations and equations n~cl c_~,y to provide corrections for such differences.
The following defines terms used in equations which calculate ratios n~r~r~RcAry to control a turn of AGV
15 2A:
C e a constant Rv = radius 84 of turn of the vehicle as the distance between turn center 82 ~nd vehicle center 86 Av = center 86 to encoder 58 distance Bv = center 86 to wheel (57 or 59) distance rc = ratio calculated using encoder 58 dimensions r,v = ratio calculated using drive wheel dimensions Wr = error due to differences in encoder 58 and wheel 57, 59 dimensions through a move (+ sign of w, signif ies 3 o direction of turn) The ratio of the turn in encoder dimensions is ~' C -- ~hc (Wr~ rR ~ Av) rc = C + abs (Wr) = (Rv + Av) WO 92/09941 PCr/US9l/0~892 -~09~442 Solving for R~, yields, I 1 + r~ I
R~, = A~ rc ¦

Thus, ~v r~ + By which is the actual control ratio o - used to power wheel drives 8, 10.
As seen in Figure 44, encoder 58 comprises a wheel 98 attached to a spring loaded AGV 2A axle 72 which retains firm but constant wheel 98 contact with the ground independent of AGV 2A load. Encoder 58 al60 comprises an Pn. o~;nq tr~nc~l7rpr 88 which provides an electrical mea~uL~ L of travel and at least one semiconductor encoder chip 90 which receives encoder 58 output, ~cumlll ates a count related to travel distance and direction, and provides output bus communicating lines to a computer plocessul. The output of Pn~orl;n~
tr~nR~lcPr 88 comprises two wave Eorms, a phase a signal 102 and a phase b signal 104, as seen in Figure 45.
Signals 102 and 104 are square waves whose phase relationship changes based upon direction of travel. An encoder chip 90, receiving signals 102 and 104, inuL- Ls (or decrements) a counter ~pppn~lpnt upon rate and direction of travel.
A simplif ied block diagram of the connection between wheel 57, 59 encoders 98 and the outerloop 3 0 processor 67 and the innerloop processor (motion control processor 61) is seen in Figure 3, 4. As seen therein, the Pn~o,l;n~J tr~nc~1llcpr 88 is connected to two encoder chips 90. To encoder 98 is thereby connected through lines 772 and lines 772A to a chip 90 which connects through lines 778A to outerloop processor 67. Similarly, bottom encoder 98 is also connec~ed through lines 776 and lines 776A to a chip 90 which connects through lines 778C
... . . . .. .. . _ . ... . .. . . . . _ . , , _ _ _ ,, WO 92/09941 2 Q 2 ~ 4 ~ 2 P~/US91/08892 to outerloop ~LucebsùL 67. In addition, the angular output of gyro 500, as described herein, is connected through lines 774 to a chip 90 and thUL-'rL~ through lines 778B to outerloop p~ oC~:s~ùL 67 . Thereby, all of the mea ul- Ls from wheels 57 and 59 and gyro 500 are made available to outerloop l Lùcessul 67.
Further, a connection is made from lines 772 through path 772B to a chip 90 and therefrom through lines 782A to motion control l-LucessoL 61. As well, a connection is made from lines 776 through path 776B to a chip 90 and therefrom through lines 782B to motion control pLU-_tSsc~ 61. These provide the inputs required for innerloop processing.
As seen in Figure 44, encoder 58 comprises a wheel 98 attached to a spring loaded AGV 2A axle 72 which retains f irm but constant wheel 98 contact with the ground i nSPp~nrq~nt of AGV 2A load. Encoder 58 also comprises an ~nro~l; n7 trAnC~ltlc~r 88 which provides an electrical mea~uLI ~ of travel and at least one semiconductor encoder chip 90 which receives encoder 58 output, accumulates a count related to travel distance and direction, and provides output bus ~ ting lines to a computer processor. The output of encoding tr;ln~ r 88 comprises two waveforms, a phase a signal 102 and a phase b signal 104, as seen in Figure 45.
Signals 102 and 104 are square waves whose phase relationship changes based upon direction of travel. An encoder chip 90, receiving signals 102 and 104, ln. L~ c (or deuL Ls) a counter ~l~p~n~l~nt upon rate 3 0 and direction of travel .
A simplified block diagram of the connection between wheel 57, 59 encoders 98 and the outerloop processor 67 and the innerloop pLUC~ssoL (motion control proces60r 61) is seen in Figure 46. As seen therein, the ~nrorl;n~ tr;~n~ulllr~r 88 is connected to two encoder chips 90. Top encoder 98 is thereby connected through lines 772 and lines 772A to a chip 90 which connects through WO 92~09941 ~ 91/08~92 2b9~ 2 ~ 116 PCI/US
lines 778A to outerloop processor 67. Similarly, bottom encoder 98 is also connected througl~ lines 776 and lines 776A to a chip 90 which connects through lines 778C to outerloop ~Lucesbu~ 67. In addition, the angular output 5 of gyro 500, as described herein, is connected through lines 774 to a chip 90 and therefrom through lines 778B
to outerloop ~LocessuL 67. Thereby, all of the mea~uLI ~s from wheels 57 and 59 and gyro 500 are made available to outerloop UL ucessùl 67 .
Further, a connp~fion is made from lines 772 through path 772B to a chip 90 and therefrom through lines 782A to motion control pLucessuL 61. As well, a connection is made from lines 776 through path 776B to a chip 90 and therefrom through lines 782B to motion 15 control ~,uces6uL 61. The6e provide the inputs required for innerloop prorPccin~.
Calculations for the U~date Marker SYStem When traveling under self-contained guidance between f loor marker 6 updates, AGV 2 continuously 2 0 searches f or an update marker 6 . As update markers in the currently preferred Pmho,l; L are magnet6, the following description will substitute descriptions of magnet 6 sensing in place of the more general update marker 6, although the invention is suf f iciently broad to 25 use update markers which are different from magnets.
As the moving vehicle 2A LL~IV~:~ es a magnet 6, the position of the magnet 6 is sensed in the vehicle 2A
frame of reference. As seen in Figure 69, the signal 403 sensed from a LLIlvt:L~ed magnet 6 results in a delayed 30 recognition of the peak of the signal at a point 403 which is offset from magnet 6 centerline 557. As well, resulting control action based upon vehicle 2A outerloop 820 adds further delays from point 403 to 464 which are ~ler~ntlPnt upon vehicle 2A velocity relative to outerloop 35 820 computational speed.
The measured parameter is the offset from the direction of travel or the "Y" offset from the vehicle _ WO 92/09941 ~ ~ 9 ~ PCr/US91/08892 center line 559 (see figure 34) of ~agnet 6. Delays in the time of actually acquiring the position after vehicle 2A traverses the point of mea_uL~ ~ generates an "offset" in addition to the actual mea~uL~ L. The 5 "offset" i8 referred to herein as latency. As an example, the ~ of vehicle 2A between sensing of data and position determination by vehicle electronics correlates to an error in the direction of travel, referred to as X~ y~ Other errors comprise errors in lO mounting which generate error6 in both the xv and yv directions, where the superscript v denote6 vehicle frame .
The X~ cy is described and estimated as follows:
X~ y = A + B (vehicle 2A speed) Where: A is a function of magnet 6 vertical alignment and other f ield abnormalities and magnet sensor and Hall sensor array 24 vehicle 2A
mounting .
s is a function of vehicle speed.
After calculating X~ cy~ vehicle 2A position is converted to factory frame 736 coordinates. X~y i5 subtracted from the estimated position of the vehicle and added to the coordinates of ~lc.v.:Laad magnet 6 and the 25 expected vehicle 2A to magnet 6 position is calculated by X I F = ¦ Cos ¢' -- S in ~ I I --X~y ¦ + ¦ X
Y I V/b I Sin ~ Cos ~ I I O ¦ I Y I V
3 O where V/M is the vehicle position when magnet sensed M/V is magnet position as calculated by vehicle M/F is magnet position defined in factory coordinates .
-- _ _, WO 9~/09941 2 ~ 9 5 4 ~ 2 PCI/US91/08892 Convert "-Y " offset to factory orientation X I f ¦ COS ~ - SIN /p ¦ ¦ o ¦ Translation of Y values ¦ Y ¦ ~. I SIN ~ COS ~ ¦ ¦ Y
Compute the position where the vehicle has det~ min~d the magnet is located. Y position is equal to the current vehicle position in factory distance moved 10 since passing magnet.
I X If I X 1~ I X If l = l l _ Y I V I Y I V~C~ I Y I Y,~
~ ~ ~ ~ ~ --Compute error di~ference XP p yP p and X~slv/ YhW
I X I l X IP l X IP
I Y l,,~ I Y IM~P I Y IY~UV
l X IP
where l l is def ined in the AGVC tables I Y I MIP and is transmitted to vehicle At beginning of next mo~e segment use X ¦ P to correct the ¦ X ¦ P
X~,~", Y"ll," ¦ Y ¦v vehicle po6ition ¦ Y ¦v X IP I X IP I X
l = l l +
I Y Iv I Y Iv I Y l.. ~
_ _ _ _ WO 92/09941 2 o 9 ~ PCr/US91/08892 V~h;~le 2~ In6ertion into FactorY r - 736 A process which uses vehicle 2A contained update marker 6 sensing to establisll an insertion ~L~J~eduLa of vehicle 2A in factory ~Erame 736 is as 5 follows. Two update markers 6 seen as 6 and 6 ' in figure 54 Are LL,~ L .ed by a manually driven vehicle 2A. The path def ined by the straight line between markers 6 and 6 ' makes an angle 756 with the factory frame. Each marker 6, 6 ' (Ml and M2) has a predel:ermined position ~ ~ ~ ~
¦ X ¦ ¦ X IF
y ¦ Ml and ¦ y ¦ M~
~ ~ ~ ~ , respectively, in factory frame 736.
The following table lists terms used in this insertion description:
F = Factory Frame 736 Ml = 1st magnet 6 in insertion sequence M~ s 2nd magnet 6 ' in insertion sequence 2 0 V = Vehicle 2A
VR = Vehicle Resting place after insertion drive I = Insertion frame 768 AN = Vehicle 2A heading relative to initial heading A HEADING
B = Angle between Ml & M2 in }nsert frame 768 XN~ YN = Vehicle 2A coordinates as calculated relative to Ml C = Angle between M~, M2 in Factory Frame 736 a " ' " indicates a measured value.
V/MI = V to Ml or vehicle to magnet 1 As vehicle 2A traverses marker 6, a measurement is made. At the time such mea-.uL~ L is available for navigational use, an offset has resulted along the line o~ vehicle 2A travel 766 moving vehicle 2A to a position Xl~ y~ In addition, manually driving vehicle 2A over update marker 6 generally result6 in a Y offset 762 as WO 92/09941 PCr/US91/08892 2~9~A~2 120 seen in Figure 54 . Thus, the ini~ i ~ l i ze~ navigational values are I x 1' I Xla~ I
5 ¦ Y I V/MI I Y762 J ~
The transformation to insertion coordinates is provided by ¦ X ¦ ' ¦ COS A --SIN A I I Xh~Y
Y I Y/M2 = ¦ SIN A COS A I I Y~2 Where, for example, Initial vehicle 2A position is defined as point X~ y past the sensed point of M~ and Y7a2 is the distance to Ml from magnet sensor in the center of the vehicle .
The angle 756 (B) is angular measure prP~lPtPrmi~pd by the locations of markers M~, M2 (6, 6').
20 Thus, for angle 756 in Figure 54:

B ~ tan~

~
Similarly, with reference to factory frame 736:

C = tan~~
3 I X~2 XMI
and, theref ore, X ¦ = ¦ cos (B--C) sin (B C) I I XVR XV/M2 1 + I XM2 Y ¦ v ¦--sin (B--C) cos (B--C) I I YVR YV/M2 I I YM2 .

WO9Z/09941 20 ~ ~ 4 4 2 PCr/US91/0~892 The vehicle 2A insertion process is as follows:
1. Drive vehicle 2A across marker 6. Measure Y distance 762.
2. Assume initial heading angle i8 l~oll and 5 position of M~ (marker 6~ in insertion frame 768 is 0, 0(x, y). Therefore, the initial conditions are X ¦ ~ ¦ Xl""q ¦ after Nl Y I V/MI = I Y762 I meaDu~ ~ L
~ ~ ~
3 . Drive (manually) across second marker 6 ' (M2). Neasure Y distance 764. Then, . ~ ~ ~
X I I I X~q 15 I Y I VIM2 = I Y76 ~ I_ The derived mea~uL Ls now allow x, y to be computed by the navigation and guidance computer.
Au2 - heading from which a calculation iB
2 o made f rom the path to the next waypoint .
4. The position of M2 in insertion frame 768 (relative to Ml) is computed by ¦ X ¦~ = ¦ COS A --SIN A ¦ . ¦ X~"q, Y I U11~2 ¦ SIN A COS A ¦ ¦ Y~64 ~ ~ J
The angle is computed by 3 0 I Yut~2 B = TANI ¦ Xul/u2 .. ~
5. Drive vehicle 2A to a rest stop 758, STOP, REQUEST INSERT FROM AGVC 13.
6. On receipt on Insert from AGVC 13 (Actual Ml, M2 coordinates) WO 9~/09941 1 = PCI/US91/08892 20=95442 YM~ YMI
C = TANI ¦ ¦ (See Figure 53 for definition XM2 - XP~ I of C.
~
Compute position in factory coordinates X IP = ¦ COS (B--C) SIN (B--C) ~ VR_XVnQ ¦ ¦ X IP
l+ l I Y I V I--SIN (B--C) COS tB--C) I I ~VR--YVna ¦ ¦ Y 1 M2 ~ ~ ~ ~ ~ J
7 . The heading of the vehicle in the f actory frame i5 calculated by 15 EEADIN~ V EEADIN~ a e (B C) WO 92/09941 2 D 9 ~ ~ 4 2 Pcr/us9l/o8892 Details of Vehicle 2A Calculations Inrl~ ; ne KA 1 r~n EquatiQns Some earlier described material is repeated in this section for clarity. The eguations for navigation 5 and g~ nne of the vehicle 2A and for the Kalman filter update of system parameters are baE~ed on aperiodic observation of ~u-v~:y~d floor position ground markers 6, aperiodic comparisons of vehicle 2A to factory frame 736 a2imuth detormlno~ by the inertial table 700 with that 10 based on fifth and sixth wheel (57, 59) data, and initial azimuth initialization estimates inserted manually into the system as herebef ore described . Derivation of the eguations are not provided. In this section, the eguation sets are rnn~iciored to pro~ide for three 15 separate computation cycle times. The first and fastest cycle is called the steering command loop. This loop provides a steering command to the vehicle steering computer and pre-processes wheel encoder data for a navigation loop. At this time it is not certain that 20 these _LItions will have to be performed at a faster time than the navigation eguations. If they do not, the eguations given are easily modified for inclusion in a navigation loop. The time for the steering loop is designated T,. The navigation loop contains the basic 25 vehicle navigation eguations, calculation of the steering command angle needed by the steering loop, and certain integrations needed in the Kalman f ilter loop . The navigation loop time is T".
The Kalman f ilter loop time is variable and is 30 designated Tl. In the currently preferred omhorli~ L, a Kalman f ilter cycle occurs immediately after the detection of a marker 6, at which time an observation is made of vehicle azimuth angle differences (table 700 vs.
wheel 57, 59 data) and upon initial insertion of manual 35 azimuth angle into the system. Each observation is ~ssumed to be in~oron~qont of the others (down-range vs.
cross-range in the vehicl~ frame because of their _ _ _ _ _ _ _ _ _ _ _ _ , _ : : . . . . .

-wO 92/09941 I PCr/US9l/08892 20~ 2 orthogonality). Thus, the Ralman filter will process marker observations (sequentially) and angle comparisons Cvll~;uLLellLly .
References of distanoe an~ angle to frames of reference are as follows:
X~ where X is the Carte6ian mea~u, ~ of x referenced from the factory to the vehicle frame. Frames used herein consist of:
Vehicle (v) Factory (f) Waypoint (w) Inertial Table (t) Others are as specif ied.
~ where ~ is an angle measured from factory ~rame to vehicle frame.
Frame references are as specified above.
In some cases, angles are measured relative to two inte~secting lines (e.g. s and ss).
Then the angle from s to ss is ~" .
The first step in Kalman calculations initiAli7~ the system error covariance matrix P as seen below, where all p~ values are taken at time (0) W(>9~l0994l 2 0 9 a 4 4 2 PCr/US9l/08892 p~ O o o o o o o o I O O p33 0 0 0 0 0 0 I Pu P~ O O O O p5~ 0 0 0 0 o O PT7 1 P..

Where Pl~ = E[ (txr)2]
P22 = Et ( ~Yf) ]
P33 = E[ (~or)2]
P~ = Et (~K,)2]
p55 = E[ (i~k,)2]
P,~6 = Et (~r.)2]
p7~ = E t ( ~ Yf) 2]
P~ = Et (~ADCV) ]
Pss = E[ (~S~RW) ]
Where the error state variable6 " ~" are def ined as 6xv _ x error in the f actory f rame ôy~ --- y error in the factory frame ôOf -- heading error of the vehicle in the factory frame ôk, _ error in right wheel calibration ~k~ _ error in left wheel calibration cr, --- error in axle (rl) calibration ~i'yf _ error in gyro drift of the inertial table relative to the f actory f rame _ error in the drift rate estimate due to Markov noise ~(iJRW _ error in the drif t rate estimate due to random walk noise WO 92/09941 PCr/US91/08892 2~95~2 The error state vector, ôX, s defined as:
l ~Xv I ~Y~
1 ~0 ¦ ~K, ~X = ¦ tK, I ~r.
I O~ v I IS~RW
The initial P matrix can be generated by performing the following calculation:
P(k) = ~p(k) (.5Q(k) ) ~T(k) + .5Q(k) where k = 0 To initialize using mea~ur I ~:j, which include typical sy6tem noise, vehicle 2A is moved between two markers 6, 6' having pre~lPtP~ninD~ factory frame 736 locations. As vehicle 2A is moved, "noise" is propagated down this "typical path," designated by ~p. Definition for Q i6 provided hereafter. As disclosed earlier, P88(0 is set equal to zero as ~larkov gyro drift error is negligible in the currently preferred P~nho~
ca~ Ations Done in the Steerinq LooP
The steering loop is the fastest loop. The steering loop time is designated as T,. Navigation and guidance loop time (TD) is currently 30 m~lliceco~
Steering loop time can be as slow as navigation and ~. i ~ .n- e loop time T", but is cyclicly tlPpPn~Pnt upon vehicle 2A guidance stabilLty requirements. In the currently preferred P"~ho~ t, T~ is an incremental multiple of T, as related below. In the ~ollowing rllcfiirn, vehicle 2A is assumed to be going forward such that starboard = right and port = left.
Input to steering loop equations:
in~;L~ I al angle (ticks) from the fifth (starboard) wheel 57 encoder.
_ _. _ F
_ _ _ _ _ _ _ WO 92/09941 2 ~ 9.~2 PCr/US9l/08892 ~h~ Lc~l angle (ticks) from the sixth (port~ wheel 59 encoder.
~c(n), steering command angle from navigation and ,~ guidance loop calculations.
Output of equations and &elective constants:
RRW Kalman correction of CD
~f(n) vehicle azimuth derived from fifth and sixth wheel (57, 59) encoder 58 data.
Sent to navigation and guidance loop at the end of "M" steering loop cycles- (Tn =
M * T,) I~X"~ (n) scaled starboard wheel 57 ticks ted over "M" steering loop cycles and sent to the navigation and guidance loop.
~x~ (n) scaled port wheel 59 ticks Arc~ Ated over "M" steering loop cycles and sent to the navigation and guidance loop.
~r = break frequency of gyro Markov noise CD = initial estimate of the gyro drift rate ~o~(s) steering command sent to vehicle steering computer every steeriDg loop cycle.
S""", = nominal scaling factors to convert right, left wheel encoders 58 SA ~ ~ 1tr, R~ h = Kalman correction of S"" ~"
Equations:
~,(5) = S~,, * ~ "(5) * (1 + R""); ~x,~(o) = O
~ ,(n) = ~ "(s) over "M" cycles (one navigation 3 o and guidance cycle) .
I~X}U(S) = S~V * AO~(S) * (1 + R~.,); ~x~(o) = O
~X~.,(n) = ~ ,(s) over "M" cycles ~Of (S) = S, * [~X",(S) - D,X",( S) ]
~f (S) = ~f (s-1) + ~Of (S) Of(n) = Of(s); fJf is read out once each navigation and guidance cycle.
~e(S) = ~c(n) ~ 0~(5) WO92JO9941 5 ~ Pcr/us9l/o8892 ~ (s) = S,~*~c(s--l) + 5.2*~c(5--2) + S~3 t1e(s) +
s,4*~(s-l) + s,5*~c(s-2) r~ l culati~nc Done in the Naviqation and Guidance LooP
The following calculations are performed in the 5 navigation and guidance loop.
TD navigation and guidance loop time (30 ms) is the currently preferred r~mho~
~ XV(n) = [~Xnv(n) + Xlw(n) ]/2 AyV~O
COS(n) = [cos~vf(n) + co60v~(n-l) ]/2 SIN(n) = [sinOvr(n) + sinflvr(n-1) ]/2 (n) = COS(n) * ~XV(n) ~(n) = SIN(n) * ~XV(n) xr(n) = xvr(n-l) + xV(n) YVr(n) = Yvf(n-l) + ~yv(n) ~ I.DCV(n) = c~v(n~l) * (1-) = O (in the currently preferred ~ t using gyro 500) L~tr(n) = ~ V(n) + CD + KRW (gyro 500 drift rate) (n) = ~y~f(n-l) + TD * ~d~f(n) ~l3(n) = ~l3(n-l) + ~y~(n) ~:,4(n) = ~4(n-l) + COS(n) * ~Xnv ~s(n) = ~:~5(n-1) + COS(n) * ~X~w (n) = ~3 (n-l ) + ~xf ( n ) ~2,.(n) = ~:24(n-l) + SIN(n) * aX,w(n) ~25(n) = ~2s(n~1) + SIN(n) * ~XIw(n) ~:3J ( n) = ~34 ( n-l ) + Xnv ( n) S.35 ( n) = ~35 (n-l) + ~Xlw ( n) (n) = ~36(n-l) + /~vf(n) The following equations are calculated at the 30 initialization of a straight line maneuver at time I.
O~, = the eath anqle to 2 way~oint 714 in factory frame 736 coordinates.
xvf(I) = Xv(n) Yvf(I) = Y'(n) x`vV(I) = [Xw - xv(I) ]Cosow + [Y~v ~ YV(I) ]sinOW
Yv(I) = [Xw - Xvf(I) ]5inQ~V ~ [Yw ~ Yv(I) ]CS~w Xo = xW(I) WO 92/09941 2 ~ 9 5 4 4 2 PCI`/US91/OX89Z

yO = y~(I) d = XO
aO = Yo r(~ w 5 al = tan,B ( I ) a3 = - ~yO - 6*tan B rI) d3 d2 a~ = 15*yO + 8*tar~ BrI) d4 d3 a5 = - 6*Y~ - 3*tan l~rI) d5 d' Xo(I) = XW(I) -- Xvw(I)co5~w Yo(I) = YW(I) - Y~(I)C05u The following e~auations are then calculated at navigation and guidance loop rate ( i . e ., at ~) xV(n) = [xV(n) - Xo(I) ]COSOw + [y~(n) - yf(I) ]8in~w yv(n) = - [Xvf(n) - Xo(I) ]Sinow + [yr(n) _ y~(I) ]Cosrtw yD(n) = aO + a, * xVW(n) + a3 * (Xvw(n) )3 + a, * (Xvw(n) )4 + a5 * (Xv (n) ) 5 ôp(n) = ydw(n) - y~(n) The bi equations are calculated at the initialization of a polar turn. The other equations are calculated at navigation and guidance loop rate. RT is the radius of the turn. Xw, Yw are defined here as XV(I), Yf(I), the initial starting point of the vehicle in the factory frame .
XP = RT + [XV(n) - XW]SinO~, ~ [Yf(n) ~ YW]C05W
-~ Yv = [Xv(n) - Xw]C05w + [yf(n) - YW]sin~w .-= tan-~ 1 xPv(n) ~ J
RPv(n) = { (XPv(n)2 + (yP(n)2~n WO 92/09941 2 0 9 ~ ~ ~ 2 PCI /US91/08892 ~

tPT = Total angle of the turn bo = RT
bz = _ T

b3 5 - ~ :
~T
b4 = BT
2*~
R~(n) = bD+bz * (~v(n) )Z+b3 * (~IpV(n) )3+b4 * (~v(n) )4 10 ~p(n) = RP~(n) - R~(n) Using either the above calculated value of ~p (n) or that derived in the straight line turn, the vehicle is steered to stay on the calculated tra~ectory using, Oc(n) = S~l*~c(n-l) + S6z*t`c(n-2) + S~3*ôp(n) + S~4*~p(n-l) + SOS* ôp ( n-2 ) (n) is used in the steering loop calculations .
~h~ RAlr~qn Filter Lool~ CalCulations ~p~3 (k) = -~13 (n) ~I~(k) = ~:l4(n) /2 ~Is(k) = ~l5(n) /2 t\z3 (k) = ~z3 (n) ~(k) = l:z4(n) /2 (k) = ~:z5 (n) /2 ~31 (k) = 5~*~3~ (n) ~35 (k) - -5.*~:35 (n) ~P36 (k) = -5.*~36 (n) ~7s(k) = [1 -exp(-~r*T~) ]/~
~7s(k) = T~
~88 (k) = exp (-~*T~,) ~,; = O (set all sums to zero) Now construct the ~ matrix.

WO 92/09941 2 0 9 5 ~ 4 2 PCr/US91/08892 .

¦ 1 0 ~3(k) ~4(k) ~5(k) 0 0 0 0 ¦ 1 ~23(k) ~24(k) ~25(k) 1 o 1 ~3~(k) ~3s(k) ~36(k) 5 ~(k)=¦0 0 0 1 0 0 0 0 0 t lo O 1 0 0 0 0 1 o o o o 1 o o o 1 o o o o o 1 ~7R (k) ~79 (k) ¦ 0 0 ~aa (k) 0 10 1o o o o o o o .
In the currently preferred ~mho~ , gyro 500 drift rate is characterized by a random walk error state variable, ~ ,. Thus, V~ = O, Vh8 = O, and hr = O, 50 ~78 =
15 ~9 = Tl" and IP88 = - T~ is the Kalman cycle time.
Cu~ r ~ion of the O Matrix The Q matrix is constructed from Q = GVGT
where G is the matrix relating process noise to the error states, and V is the process noise covariance 2 o matrix . Elements of the G matrix are contained in the equations below:
q~ (k) = v"* (~4 (k) ) 2 + v22* ('~p~5 (k) ) 2 + v33* (~,3 (k) ) 2 q~2(k) = v~*~p~4(k) *024(k) + V22*~ls(k) *~2s(k) +
v33*~,3 (k) *~23 (k) ql3(k) = v~*~4(k)*~34(k) + V22*~ls(k)*~3s(k) q2l(k) = Vll*~14(k)*~24(k) + V22*~,5(k)*~25(k) +
v33*~3 (k) *~23 (k) q22 (k) = v~* (~24 (k) ) 2 + v22* (~25 (k) ) 2 + v33* ('P23 (k) ) 2 q23 (k) = v~*~24 (k) *~34 (k) + V22*~2s (k) *~35 (k) 3 0 q3~ (k) = v~*~l4 (k) *~34 (k) + V22*~t~s (k) *~35 (k) q32 (k) = v~*~24 (k) *v34 (k) + V22*~2s (k) *~35 (k) q33 (k) = v~* (~34 (k) ) 2 + vn* (~35 (k) ) 2 - qn(k) = v7, ql~ (k) = Vsa q99(k) = v99*T
. ~_ .

WO 92/09941 ~ . PC~r/US91/08892 2~95~42` 132 The uncorrelated process noi6es are represented by the covariance matriY:
V,l O O O O O

V= I O O V33 0 0 0 O O O ~ ,0 V99 1 -- I
The values Vc in the above matri~ providc the V process noise covariance matrix.
v~ = E [ ~ ] = E [ ~2~ ]
V22 = E [ ~2~, ]
15 V33 = E[e2,] + E[~wlv]
v7~ = E [ ~WN~ ]
V88 ~ E ¦ ({2a Cb~CV4}ll2 ~ cy~2 I= 2aaMI;v 20 V99 = E[~2w]
nhere noi6e variances are due to:
floor ~ R (right wheel) tVil) floor anomalies (left wheel) (V22) wheels side slip (V33) white noise in angular rate sensor (V~) Narkov noise in angular rate sensor (MKV) where a~ V i5 the standard deviation of the Markov nois= (V~) WO 92/09941 2 0 9 5 ~ 4 2 PCI /US91/08892 random walk noise is angular rate sensor (V99) ql, (k) q~2 (k) ql3 (k) 0 0 0 0 0 0 ^ I q2~(k) q22(k) q33(k) 0 o o o o o 5 ¦ q3l(k) q32(k) q33lk) 0 0 0 0 0 0 Q(k) =¦
O O O o O O O o O
l O O O O O O O O O
O O qT7 (k) 0 ¦ 0 0 0 0 0 0 o q88 (k) o o o o o o o o o q99(k) Next, propagate the error covariance matrix P
to time k using T~, the time since the la6t Kalman cycle.
P(k) = ~(k) [P(k-1) + lt2*Q(k) ]~(k)r + 1/2*Q(k) Now, construct the h vector (s) appropriate to the observation being processed.
h~ = [cosO~sin~t o 0 0 0 0 0 0]
h2 = ~-sin~rcosOf o o o o o o o]
h3 = [0 o 1 0 0 0 -1 o 0]
In the above equations, all mea~uL~ Ls are assumed to be i n~PrPn~3Pnt~ providing reclllnA~nry whereby systematic errors of mea~u~ L are evaluated and removed by real time calibration using Kalman f iltering .
25 ~larker 6 mea2,uL~ ~s (x, y) are used in observations 1,2 and vehicle 2A gyro 500 mea~uL~ L is used in observation 3 of the h vector(s), above. The down-range and cross-range observations are orthogonal in vehicle 2A
coordinates and could be measured at different times. of 30 course x and y should be measured ~,Ull~;ULL~IlLly. However, in the currently preferred Pmho~i - t, a sample directional mea~uL~ L is taken from gyro 500 each time a marker 6 is located. In another preferred Pmho~l;r L, a sample directional mea~uL L is also taken from gyro 35 500 each time a new turn or straight line path is calculated . Thus, the h matrix comprises hl 2 3 in the currently pref erred ~ '~o~ nt .
~ = , .

WO 92~09941 PCr/US91/08892 2ass442 Note: While all measurements are ~ r~n~l~nt, the down-range and cross-range observa~ions are orthn~r~nAl in vehicle coordinates, and can be taken separately, but x and y 6hould be taken simultaneously. ~, Notation: Kj, hj, rj, Zj, . . .; i = the measurement number i. e., i = 1 down-range observation i = 2 cross-range obse~vation i = 3 azimuth angle comparison The Kalman gain Ki for the i'th observation where rj is def ined as the i ' th observation noise variance is computed.
Ki = P(k~hT
(hjP (k) hT +
The f ilter covariance matrix is updated by the f ollowing equations:
P+(k) = (I - K,hj) = P(k) - Kj(P(k)hT)T
where P+ (k) is updated P tk) ~Zj = hj~X
~ X~ = ~X + Ki(~Zi - ~Zi) Where " ^ " over a variable denotes the estimated value of that variable. The initial estimate for ~X is zero at th~he~innin~ of each Kalman cycle.
For simultaneous observations, such as observations 1, 2, an iteration process is perf ormed as f ollows:
Kl = P(k)h h~P (k) hl + r ôz~ = hlôX = O
~S~f = kl~zl ., 6X = ~X+
P+ (k) = P (k) - Kl (P (k) hT) T
P (k) = P+ (k) K2 = P (k) h2 h2P (k) hT + r2 WO 92/09941 2 0 9 ~ ~ ~ 2 PCI /US91~08892 135; ..
~Z2 = h2ôx ~x~ = ~x + k2(ôz~ - ~SZ2) P+(k) = P(k) - k2(P(k)hlT)T
P (k) = P+ (k) The final values in ôX+ provide the state variable corrections, as error correcting estimates.
Review of th~ R;llr-n Filter as U8ed ;n th~ rrently Preferred Embodiment Process errors are ba6ed on specif ications, drawings and calculations . Once the process noises ( 6 maximum) and mea~uL ~ t_ noises ( 3 maximum) are selected, then the ~r u~oyu~ion of process and mea~u . ~ noises is def ined by the path ( s ) .
Process Noises ~
¦ V~ ¦ <--Floor, wheel irregularities V22 ¦ <--Floor, wheel irregularities V33 ¦ <--S ides l ip V= I O

V66 ¦ <--Gyro Random Walk J
Mea~ u, ~ Noise ¦ r~ o o ¦ <--Magnet position & heading calculation from encoder 58 R = ¦ r22 ¦ <--Magnet position & heading calculation from encoder 58 , r33 l <--Gyro 500 encoder resolution ~ ~ & Heading calculation from encoder 58 To estimate the f irst initial P matrix, one can use WO 92/09941 PCr/US91/08892 ~95~ 136 O O O O o l O O O O O
I O O O O O
P~ (k) (.5Q(k))~(k)T+ 5Q(k)+ ¦ O O E(~kr)Z O
¦ O O O E(~k,)20 l l l O O O O O
where k = O
10 in which E(ôkr)2 = expected value of the squared error in wheel diameter ( ô`k2, = ~Sk,2) NOTE: ~ and Q are based on a "typical"
path/pathlength .
It was previously shown that a marker (magnet) observation required two measurement inputs related to two of the error state variables (ôxv', sy,f,). With error states ~r" 6 yf~ CV removed from the error state vector, the def inition of the measurement can be shown to be ôz = h~X
(2 x l 2 x 6 6 x l matrix orders) ~Meas . = hô state where ~z = detected - calculated mea"uL~ Ls ôX = actual - estimated (calculated) states - . h is selection vector.
The "pref erred ~ ' i L" involves ~ taking an azimuth mea_uL L at the ti--- the magnet is detected to provide overlapping, redundant information.
The azimuth meat~u.. ~ by itself is actually 3 o given by:
ôZ3 = O heading error - ô gyro drift error ~; ~ f _ T~ f encoder 58 Kalman Drift rate wheels 57, time error 59 error _ -- -- -- = F

WO 92~09941 2 ~ 9 5 -1 4 2 Pcr/us9l/o8892 Careful expansion shows that = T~ + ~' - of Predicted Change in Gyro Encoder 58 Drift ba6ed 500 Encoder for wheels on last 57, 59 t~df) updated drif t rateO is equivalent to ~Z3 = (0 o 1 0 0 -T~) ~X
where ~t is the change in gyro 500 angle over the last Ralman cycle i5 the current estimate for the underlying5 variable (e.g., Of iS the current estimate of 0 measured from factory to vehicle coordinates. ) Using a simultaneous update actually reduces process time from a sequential method, besides offering 20 the advantages of red~ln~lAnry~ "best information, "
stability, etc. 50 that the vehicle can operate at up to 200 ft/minute with no need to slow down for Kalman calculations .
To maintain speed, Kalman updates are onlv 25 performed before a new trajectory calculations at a waypoint or after sensing a marker 6. An azimuth mea~uL, ~ can be performed prior to any waypoint. To combine azimuth and marker mea~uL~ Ls, the mea~uL~ L
matriY h is Dyr~nrlDd to be used as follows:

WO 92/09941 . ~ . PCr/US91/08892 209~442 I I ~ I I ~xr -~sin~f ~ ~ I l l O ~COSOV ¦ ¦ ~Y~ ¦
l l 11 ô5 ¦ ~ ~ ¦ ~ kr ~cosOv ~ ~y l=1 1 ~5inOf o o o l l ~k~ ¦
I ~ ~ I I 11 ô~f I
¦ T~ ( O ) + ~'YV - v ¦ ¦ O O 1 0 0 -T}
I ~
1 0 ~ J
(ôz = hôX) where c~(O) is the value of ~ at time t(o) where ~ is yv position acquired from a marker 6 mea,,~L~ L and ~ = Xv ~ ~
where "m" is the measured marker position ~y YV ~ Y~
~ During that Kalman time [T~ = t(1) - t(O) ]
the transition matrix, ~, is propagated as shown earlier.
At time t(1), the P matrix is propogated as follows:
P (k) = ~ (k) (P (k-1~ + . 5~Q (k) ) ~ (k) T + . 5~2 (k) then K = P(k)hT(hp(k)hT + R)-l Kalman gain ~+ = K~ôzz Updated errors [actual -calculated estimated]
then when the waypoint is reached where X(n) is the current estimate X(n) = X(n-1) + ~X+ ¦ of the state of the vehicle parameters ( i . e ., ~ = tXv Yv ~Uv Kw Klw ~-] ) P+ (k) = (I - K h) P (k) P (k) = P+ (k) _ .

WO 92/09941 ~2 ~ 9 ~ ~ 4 2 PCr/US91/08892 V~h i cle ? A Path Selectinn Criteria A6 described earlier, each calculated path between waypoints 714 is based upon the calculation of coefficients for a ~ifth order cartesian coordinate or 5 polar coordinate polynomial. Each polynomial being used for calculating in 6eriatim the guide points of a path along which vehicle 2A steers, the coefficients of the cartesian coordinate polynomial being used for "straight line" moves and the polar coordinate polynomial being 10 used to calculate moves involving arcs, such as paths around corners. In some cases, both straight line and arc moves are used in sequential combination in those circumstances where a single move will not place the vehicle at the target waypoint 714 procDe~lin~ in the 15 desired direction.
Polynomial yeneration constraints f or the linear or "straight line" coefficient calculation comprise restricting the initial angle of the path to the then current direction o~ AGV 2A, defining a path which 20 is continuous, and terminating the path at the target waypoint 714 tangent to the heading specified by AGVC 13.
Figure 64 provides a graph of a desired path between the vehicle and a next waypoint 714. The vehicle position 710 and direction 738 are usually not as originally 25 planned and stored for the path between waypoints 714.
Instead, in realistic terms, AGV 2A must be considered to reside at position 710 when the 710, 714 path calculations are made. Constraints placed upon calculation of the coefficients of the fifth order 3 0 polynomial which describes path 716 in Figure 64 are the following:
1. Initial heading of the path 710, 714 is equal to vehicle 2A heading 738.
2. The path described by the polynomial is 3 5 continuous .

WO 92/09941 2 ~ 9 5 ~ 4 ~ PCr/US91/08892 3. Final heading of vehicle 2A at 714 is the heading angle provided by AGVC 13 in the message directing vehicle 2A travel and is always zero in waypoint frame 734.
5 Exemplary path 716 shows expected ;UL va~ur ~ in the "straight line" guidepath.
Dc ~ dl=--t upon vehicle position relative to a target waypoint 714, one set of polynomial coefficients i5 generated. A new set of polynomial coefficients is 10 generated for each different position of vehicle 2A
relative to waypoint 714, thereby providing a family of curves, three of which are seen in Fiyure 66. As seen in Figure 66, guidepaths 716A, 716B, and 716C provide three such exemplary guidepaths f or a vehicle placed at an initial position 710 and heading 738 near a first waypoint 714A.
Polynomial generation constraints for the polar or curve guidepath are similar to those of the "straight line" case, resulting in an arcing path as seen in Figure 65. The constraints placed upon ~he coefficients of the fifth order polynomial in polar coordinates are as f ollows:
1. Initial heading of path 726A is tangent to the initial heading of AGV 2A at the beginning of 2 5 - the path .
2. The fifth order polynomial describes a continuous path.
3. Final heading of vehicle 2A at 714 is the heading angle provided by AGVC 13 in the message directing vehicle 2A travel and is always zero in waypoint frame 734.
Exemplary path 726A, seen in Figure 65, shows the expected guidepath arc along path 710, 716.

W092/09941 ao ~5 44~ - PCr/US91/08892 D~pen~ nt upon vehicle position relative to an initial waypoint 714A, one set of polar polynomial coefficients is generated. A new set of polynomial coefficients i8 generated for each different position of 5 vehicle 2A relative to waypoint 714, thereby providing a family of curve5.
It has been found in the current ~mhor~ nt, that selection of the type and number of polynomial calculations to define a guidepath is ~l~r~n~l~nt upon lo initial heading 738 of AGV 2A in waypoint frame 734. If the initial heading 738 in waypoint frame 734 produces an angle with the i~lhC~i Ccn of waypoint frame 734 which is an absolute angle less than 20, a "6traight line"
calculation is made in Cartesian coordinates. If the heading angle produce6 an angle with the AhCcicca o~
waypoint frame 734 which i5 greater than or equal to 20, a decision process, best described by referencing Figure 96, is ~ollowed. As seen in Figure 56, initial heading 738 results in angle 1210 with the ~hsl ic~ca of waypoint frame 734.
The next step in the decision process calculates a polar or curved path 1240 which tangentially intersects the i~hcc; Csa of waypoint frame 734 at a point 1250. The intersection 1250 is a distance 1230 from the origin of waypoint frame 734. The distance along the scicc i of waypoint frame 734 to waypoint 714 is seen as line 1220. If the length of line 1230 is less than line 1220, a polar or arc move followed by a straight line move is performed. If line 1230 is equal to line 1220, a single polar move is performed. If line 1230 is greater than line 1220, and if angle 1210 is greater than 65 a ~_ ~ move, as described hereafter, is performed.
Another selecting criteria, is the use of X,Y
and the heading angle 1210 in waypoint frame 734 coordinates. The ratio of X/Y and tangent of heading angle 1210 is calculated. If heading angle 1210 is between 20 and 65, and the ratio is loss than . 35 a WO 92/09941 PCr/US91/08892 2ng~i4~2 142 "straight line" move i5 made. A family of three arch curves are seen in Figure 67. The ratio provides an estimate of the v~ hovL of the vellicle and the abscissa of waypoint frame 734 . If the ratio is greater than . 35 5 and the heading angle is between 20 and 65, a "straight line" path followed by a polar move is performed.
For those cases where the absolute value of angle 1210 is greater than 65 and line 1230 is greater than line 1220, as seen in Figure 96, the guidepath is lo defined as a combination of "straight line" and curve moves. A guidepath is seen in Figure 68 wherein the required waypoint 714 target position and direction are not achievable using a single fifth order polynomial calculated guidepath. In the case of the path 15 requirements seen in Figure 68, the path is divided into two segments, 716 and 726, the calculated path defining coefficients being calculated in Cartesian coordinates for first path 716 and in polar coordinates for second path 726, thereby achieving the n~-cQ5~:~ry waypoint 714 20 position and exit heading.
The calculation of distances in waypoint frame 734 coordinates from factory frame 736 coordinates is:
I X Iw = ICos 0 Sin l IXD - Xl I
I Y Iv l -sin 0 Cos ¦ IYD ~ Yl Where: X,Y are distances in waypoint frame 7 3 4 coordinates .
0 is the angle between the vehicle 2A and waypoint frame 734.
XD~YD are destination or waypoint 714 coordinates in the factory frame 736.
Xl,Y~ are coordinates of initial vehicle 2A position in factory frame _ 736 coordinates.

WO92/09941 2~ $ ~ 4 4 ~ Pcr/usgl/08892 For precision in calculation of positions in the currently pref erred ~ i r ~ of a marker 6 in the factory frame 736, a three byte field is used in all position coordinate determinations yielding a range of 8,368,~07 in units of 1/20 of an inch. Thus, the position measuL- 1, range i8 34,952 feet or 10,653.5 meters . A two byte f ield is used f or heading clPt~rm;n:~tions, whereby a precision of 0-359.99 is achieved for calculational purposes.

WO 92/0994~ ~ O ~ ~ ~ 4 ~ ~PCr~US91/08892 AGV 2A Central ProcP-~;n~ TJn;ts AGV 2A i5 controlled by a plurality of micro-~JL u~eSSUl r~ as seen in Figure 77 . Outerloop processor 67 is the master controller, directly ; ~-ating acro6s -bus 816 to an output ~Luces~uL 1166, an encoder processor 1164, an analog input yLucessoL 1170, the motion control p~ucessuI 61, a serial ItO communications processor 1192 and the central processing unit 810 ' for SDLC
communications chip 812. Bus 816 is referenced by other numerical identifiers within this disclosure; however, it may be c~nc;~lpred that all direct bus communications among the processors listed above traverse bus 816. Each of the processors fill an important mission for each AGV
2A. As an example, outerloop processor 67 performs the Kalman calculations in addition to other outerloop calculations and general control of AGV 2A. Outerloop ocessùl is a 186/03 board, commercially available from Intel Corporation.
ûutput pL-~ce6soL 1166 comprises pLUyLCII...~ which turn on and off AGV 2A lights, change speaker tone and duration, and actuate the AGV 2A beeper. In the current ~ ' _';r ~, output proces80r 1166 is preferably an 8742 central processing unit, now available from Intel Corporation .
Input pLocesscI 1168 provides general procPcc;n~ of input lines such as inputs from vehicle sensors, other than antennas and update markers. As an example, limit switches and emergency stop apparatus provide input signals processed by input processor 1168.
In addition, digital discrete inputs which are directed to the main ~-ùcessuL, such as 6ignals from discrete digital devices are buffered (temporarily stored in memory), then ~Luce~ed by input pr UUt:SSOL 1168. In the current Pmho(1;- L, input processor 1168 is preferably an 8742 central procPCF;n~ unit, now available from Intel Corporation .
Encoder processor 1164 processes encoder = ~
.. .. _. ... . ___.. : .:,. =___ _ w092~09g4l 20~44a' - PCr/US91/08892 signals which provide mea~,uL~ ~ information for outerloop ~Lùc~bs~L 67, such as signal$ from gyro 500, fifth wheel 57, and sixth wheel 59 travel information as used in Kalman filter calculations. In the currently 5 preferred ~ -; t, encoder proce6sor comprise6 four input ~h Innplc~ three of which are u6ed for the angular and linear travel mea:.uL L6. The fourth channel i6 a spare. Encoder plucebbuL 1164 i6 preferably an 8742 central proces$ing unit, available from Intel 10 Corporation.
Analog input EJL OCe55UI 1170 provide6 analog to digital input proc~cc;n~ wherein analog voltage input6 from each tachometer 33, 6ee Figure 4A, are received, digitized, and monitored, thereby providing a 6afety 15 backup to operation of motion control ~LU~_S6UL 61. In addition, analog input ~rOCe66uL 1170 receive6 and ce6se6 input6 from a joy 6tick on a manual vehicle control box whereby each vehicle 2A i6 manually controllable. Further, analog input processor 1170 receive6 and proces6e6 input6 from obstacle detector6 and AGV 2A battery voltage. Analog input processor 1170 is an 8742, available from Intel Corporation.
Motion control processor 61 function and r~Cp~nC; hility are described in detail earlier. As seen 25 in Figure 77, motion control processor 61 receives inputs from encoder6 58 a6 earlier de6cribed and provides digital to analog outputs which control operation of drive wheels 8, 10. Motion control proce660r 61 also provide6 a controlled "E" 6top to bring AGV 2A to re6t in 3 0 a rapid, but not hard-braking 6top in a detected U~ ,y. Motion control proce660r 61 i6 preferably a DS5000 central proc~C6in~ unit, available from Dalla6 Semiconductor. A 6econd, more direct, but gated signal path 1184, 1162, 1186 provide6 direct feedback from analog input ~Lu~65uL 1170 to motion control proce660r 61 .
In the currently preferred ~ ir~nt, output ,~ , . ~

WO 92~09941 2 û ~ 5 4 4 2 ~ PCI/US91/08892 ~

processor 1166, encoder ~Luce~ UL 1164, analog input proce6sor 1170, gated interface 1162 between analog input processor and motion control ~LUU~CSSUL 61, motion control pLU~55Vl 61, and input ~LUcessù~ 61 are installed on a 5 single digital I/O board 1194. De~ailed circuit schematics of digital I/O board llg4 are provided for completeness of disclosure in Figures 87-91, 92, 93, and 94-95. Figure 86 provides a map showing relative orientation of Figures 92 and 93. All - ^-lts seen in 10 the above referenced figures are commercially available and are used in a manner which is known in the art.
Power supply interconnections are removed for clarity of presentation. A list of component types and values, where applicable, for digital I/O board 1194 is found in 15 the following table.
1~ ~m~ Value or TYI~e UlE Optical Sw.
U2E HPRI/Bin 74LS148 U3E Latch 74LS373 20 U4E Inv.Amp. 74LS04 U5E ~Yr5~nrl-~r P8243 U6E Amplifier 74LS125 U7E Nand 74LS03 U8E Clk Gen/Dr 8284 25 U9E Octal Buffer 74LS540 UlOE CPU D8742 UllE CtrDivl6 74LS393 U12E Octal Buffer 74LS540 U13E Octal Buffer 74LS540 30 U14E Inv.Amp. 74LS04 U16E Amplifier 74LS125 U17E Comp. 74ALS521 35 U19E Inv.Amp. 74LS04 U20E Nand 74LS02 U21E Comp. 74ALS521 ... , . '' WO92/09941 ~ ~944a ~ PCI'/US91/08892 U22E F~An~ ~ P8243 U23E NultiVibtrs 74LS122 U24E OR Gate 74LS32 U25E Inv.Amp. 74LS04 5 U26E Octal Buffer 74LS540 U27E 16x4 Duel ported ram AM297 05A
U28E Octal Buffer 74LS540 U29E Latch 74LS374 10 U30E Phase Decoder HCTL2000 U31E Nand 74LS37 U32E BinlOct 74LS138 U33E Oct 8us Trscvr P8287 U34E Phase Decoder HCTL2000 15 U35E Phase Decoder HCTL2000 37E Nor LS132 U38E Phase Decoder HCTL2000 U39E Inv.Amp. 74LS04 20 U40E OR Gate 74LS32 U42E Diff.Amp. LM101AJ

U44E A/D Converter ADC0816 25 U45E Diff.Amp. L2~lOlAJ

U47E Dif ~ . Amp . LM12 4 U48E D/A Converter DAC0830 U49E D/A Converter DAC0830 30 U50E D/A Converter DAC0830 U51E D/A Converter DAC0830 U53E Phase Decoder HCTL2000 U54E OR Gate 74LS32 U55E Bin/Oct 74LS138 35 U57E Phase Decoder HCTL2000 YlE Oscillator 24 NHertz . . _ . .

- - =
WO 92~09941 ' PCr/US91/08892 ~3~42 ~

RlE Resistor 2 . 2K Ohm6 R2E Resistor lOK
R3E Resistor lK "
R4E Resistor 620 "
5 R5E Resistor 620 "
R7E Resistor 510 "
R8E Resistor 510 "
R9E Resistor 620 "
RlOE Resistor 620 "
10 RllE Resistor 30K "
R12E Resistor lK "
R13E Resistor lOK "
R14E Resistor 2 . 2K "
R15E Resistor 620 15 R16E Resistor 620 "
R17E Resistor 620 "
Rl8E Resistor 620 "
R19E Resistor 20K 1% "
R21E Resistor 7.5K "
20 R22E Resistor 20K 1% "
R23E Resistor 7 . 5K "
R24E Resistor 20K 1% "
R25E Resistor lK "
R26E Resistor 9 . O9Kl "
25 R27E Resistor 2 . 2K "
R28E Resistor 2 . 2K "
R29E Resistor lOK "
R30E Resistor lOK
ClE Capacitor 1 ~LF
30 C3E Capacitor .01 ,uF

Cl9E Capacitor lOO ,ILF
C29E Capacitor lO ,uF
C34E Capacitor lOO pF
C35E Capacitor 33 pF
35 C36E Capacitor 33 pF
C38E Capacitor lO ~F

WO 92~09941 2119 5~ 4 2 PCr/US91/08892 C39E Capacitor 100 pF
C40E Capacitor 10 ~LF
C41E Capacitor 100 pF
C44E Capacitor 20 pF
5 C45E Capacitor 20 pF
C46E Capacitor 20 pF
C47E Capacitor 20 pF
C48E Capacitor 100 ~F
ElE Jumper 10 E2E ~ ~ Jumper E3 E Jumper E5E Jumper E6E Jumper E7E Jumper 15 E8E Jumper CRlE Diode lN914 CR2E Diode L~336BZ
CR3E Diode ~329BZ
CR4E Diode lN914 20 VRlE Diode lN4733 VR2E Diode lN4733 VR3 E Diode lN4 7 3 3 VR4E Diode lN4733 VR5E Diode lN4733 25 VR6E Diode lN4733 VR7E Diode lN47 3 3 VR8E Diode lN4733 VR9E Diode lN4733 VRlOE Diode lN4733 30 VRllE Diode lN4733 VR12E Diode lN4733 VR13E Diode lN4733 VR14E Diode lN4733 VRl 5E D iode lN4 7 3 3 3 5 VRl 6E ~ Diode lN4 7 3 3 , . .,,. , ,,,,,_ .

WO 92/09941 PCr/US9l/08892 2~9S~42 150 * * * * *
Two central processing units are installed on the vehicle 2A _ i ~ ations board 824 . As seen Figure 77, communications board 824 comprises serial I/0 ications pluces~uL 1192, SDLC ;r~tions uL 810', SDI,C chip 812, and radio data recorder 820, and all related interfacing logic and other - Ls.
Serial I/0 communication ~Lu~essuL 1192 provides a ; cAtions interface for all AGV 2A
communications except for intl:~. ;cations between AGVC 13 and AGV 2A. serial I/0 com.munication6 processor 1192 also comprises an interface 1176 to update marker system ~LucessuL 482, wherefrom update marker data, processed as earlier described, is received or transferred to outerloop ~JL u~ ~88uL 67. In the currently preferred ~mho~; t, serial I/0 _ ;cations processor is a DS5000, available from Dallas Semiconductor. Update marker system 482 is a DS5000 central prQc-^ss;n~ unit available from Dallas Semiconductor.
The components and function of SDLC central processing unit tCPU 810 ' ) is described in detail earlier. CPU 810 ' is preferably an 8742 central processing unit, available from Intel Corporation.
Detailed schematics of communications board 824 are seen in Figures 78-85. All ~ ts are commercially available and are used in a manner which is known and in the art. Note that some earlier ~ ^loced circuits are repeated therein. For example, Figure 85 comprises components and circuits found in radio data recorder 820, earlier seen in Figure 74. Power supply interconnections are removed for clarity of presentation.
A list of _ ?-lt types and values, where appropriate, for communications board 824 is found in the following table.
..

-W092/09941 2 ~ ~ ~ 4 ~ 2 PCr/US91/08892 Number 1~ Value or Tv~ ~e UlD Diff.Amp. TL072 U2D Diff.Amp. IM339 U3D Amplifier DS1489 7~4D CtrDivl6 74LS393 U5D CtrDivl6 74LS393 U6D Clk Gen. /Dr. 8284 U7D Nand 74L500 U8D Inv.Amp. 74LS04 UlOD Amplifier 74LS125 UllD Power Reg. L7.~317T
U12D Inv.Amp. DS1488 U13D Nand 74LS00 15 U14D Nand 74LS02 U15D Drvr/Rcvr 75179B
U16D Inv.Amp. 74LS04 U17D Nor 74LS32 U18D Comp. 74ALS521 20 U19D Amplifier 3486 7J20D Amplifier 3487 U22D Bin/Oct 74LS138 U23D Comp. 74ALS521 25 U24D Inv.Amp. 74LS04 U25D SDLC chip 82530 ~27D Nand 74LS132 U28D OR Gate 74LS32 30 U29D Oct Bus Trscvr 28287 U30D SDLC Comm. 8273 U32D Inv.Amp. 7406 U33D Counter 74HC4040 35 U34D Inv.Amp. 7406 7J35D Buffer 74LS373 U36D Nand 74LS03 .
.

WO 92109941 PCr/US91/08892 2095~2 U37D Nand LS132 U38D ~PRT/Bin 74LS148 U39D Amplifier 74LSl25 U40D Bin/Oct LSl23 U4 lD CtrDivl6 7 4LS3 9 3 U43D Latch 74LS374 U44D Eprora 27128 U45D D Flip Flop 74LS74 10 U46D Inv.Amp. 74LS04 U47D Nand 74LS00 U48D B-Bit D/A DAC0808 U49D Diff.Amp. LE347 U50D Diff.Amp. LE347 15 U5 lD D if f . Amp . LHOD2 lCg U52D Diff.Amp. LH0021CK
Ql Transistor 2N3904 Q2 Transistor 2N2222 YlD oscillator 24 NHertz 2 0 Y2D Oscillator 12 ~Hertz Y3D Oscillator 4 . 9152 NHZ
ElD Switch SPDT
E2D Switch SPDT
E3D Switch SPDT
25 E4D Switch SPDT
E5D Sw~ tch SPDT
E6D switch SPDT
E7D switch SPDT
E13D switch SPDT
3 0 TPlD Status Ind . LED
TP2D Status Ind. LED

TP3D Status Ind. LED
TP4D Status Ind. LED
TP5D Statu6 Ind. LED

WO 92/09941 2 0 ~ 5 4 ~ 2 PCr/US91/08892 TP6D Status Ind. LED
TP7D Status Ind. LED
TP8D Status Ind. LED
,, TPlOD Status Ind. LED
5 TP12D Status Ind. LED
TP13D Status Ind. LED
ClD Capacitor 100 pF
C2D Capacitor . 001 ,ILF
C3D Capacitor . 001 ~F
C6D Capacitor .1 Uf C7D Capacitor . 01 ~F
C17D Capacitor 33 pF
C18D Capacitor 33 pF
C26D Capacitor 15 ~F
C32D Capacitor .01 ~LF
C35D Capacitor . 01 ,~F
C36D Capacitor . 01 ,~LF
C37D Capacitor . 01 ,ILF
C40D Capacitor .1 I~F
C41D Capacitor . 01 ~F
C42D Capacitor . 01 ~F
C43D Capacitor 47 pF
C44D Capacitor . 01 ~F
C49D Capacitor 3000 pF
C51D Capacitor . ~ ,uF
C52D Capacitor 4700 pF
C53D Capacitor 3000 pF
C54D Capacitor .1 ,uF
CRlD Diode lN914 CR3D Diode IN914 RlD Resistor lOK Ohm6 R2D Resistor 3 . 3K
R3D Resistor lOR
R4D Resistor lOK "
., WO 92/09941 PCr/US9l/08892 2~95~2 R5D Resistor 5. lK "
R6D Re6istor 560K
R7D Resistor 30K
R8D Resistor lOK "
5 R9D Resistor 3 . 3K "
RlOD Resistor 510 "
RllD Resistor 510 R12D Resistor 2 . 2K "
R13D Resistor 2 . 2K
10 R14D Resistor 100 "
R15D Resistor 2 . 2K
R16D Resistor 2 . 2K
R17D Resistor 620 "
R18D Resistor 620 "
15 R19D Resistor 2 . 2K "
R20D Resistor 2 . 2K "
R21D Resistor lK "
R22D Resistor 100 R23D Resistor 2 . 2K
20 R24D Resistor 2.2K
R25D Resistor 100 "
R26D Resistor 2.2K "
R27D Resistor 2 . 2K "
R28D Resistor 2 . 2K "
25 R29D Var. Resistor lOK "
R30D Resistor 100 "
R31D Resistor lOK "
R32D Resistor 510 "
R33D Resistor 1. 3K 1% "
30 R34D Resistor 100 "
R35D Resistor 4.3K "

R36D Resistor 5. lK
R37D Resistor 3 . 3K "
R38D Resistor 30K "
35 R39D Resistor 200 1% "
R40D Resistor 5K "
R41D Resistor 3 . 3K

WO92~09941 ~,0~ a'~ PCI/US91/08892 R42D Resistor lOK "
R43D Resistor 12R
R44D Resistor 3 . 3X
,, R45D Resistor 270 "
5 R46D Resistor 270 "
R47D Resistor 5. 6K
R48D Resistor 1.5K
R49D Resistor 2K "
R50D Resistor 2 . 7K "
10 R51D Resistor lOK "
R52D Resistor 2K "
R53D Resistor lOK "
R54D Resistor 5 . lK "
R55D Resistor lOK "
15 R56D Resistor 5 . 6K "
R57D Resistor 1 "
R58D Resistor lOK "
R59D Resi~tor lR
R60D Resistor lK "
20 R61D Resistor lOK "
R62D Resistor 5 . 6K "
R63D Resistor lK "
R64D Resistor lOK
R65D Resistor 7 . 5K "
R66D Resistor lK "
R67D Resistor 680 "
R68D Resistor 470 R69D Resistor lK "
R70D Resistor 75K "
R71D Resistor lOK "
R72D Resistor 2.7K "
R73D Resistor lK "
R74D Resistor lK "
R75D Resistor 2K "
R77D Resistor 2K "
R76D Resistor lOK "
. __ W0~2/09941 ao ~44~ PCr/US91/08892 A listing of pLvyLa~ used in each of the above described AGV 2A central proco~in~ units is provided in a table, entitled summary of AGV 2A Software, found below. Computer l~n~A~o~ used comprise Intel C-86 and 5 Intel As6embler in the outerloop EJL UCe~UL, other languages used are seen in the AGV 2A Software table.
The operating system f or outerloop L~L u~.es8uL 67 is the AMX86, revision 1.0 from Kadak Products, Ltd., Vancouver, B.C., Canada (copyright 1983, 1984). The AMX
10 operating system ha6 been modif ied by the inventors to support the 8087 ~u p~ucessuL in the multi-tasking environment and to provide easier interfacing between the "C" language compiler and AMX86. Source of the AMX86 compiler, as adapted by the inventors, is not included 15 because prior ayL I Ls between Eaton Kenway and Kadak Products, Inc. prohibit Eaton Kenway from publishing any AMX86 source code. Eaton Kenway can, only upon notification from Kadak that the requester is licensed and thereby qualif ied by Kadak to have access to the 20 AMX86 system, provide a copy of the modified AMX86 code.
Following the AGV 2A Software table, a listing of each software module used in the currently preferred ~-~o~i--- L is provided. The listings are paginated as indicated in the AGV 2A Software table. Also found in 25 the AGV 2A Software table are file names, file types, assemblers or compilers used (if ~rPl ic~hle), and the basic function of the software.

WO 92/09941 2 0 ~ 5 ~, 4 2 PCl`/US91/08892 AGV 2A Software File File Assembler/
Name Tv~e Com~iler Flln~tion . ANIN.ASM P 8742 ANALOG INPUT PROCESSOR 1170 CNT.ASM P 8742 ENCODER PROCESSOR 1164 COM.ASM P 8742 SDLC PROCESSOR 810 ' DSCOMM. C P C SERIAL I/O COMMUNICATIONS

DIN.ASM P 8742 INPUT PROCESSOR 1168 DOUT.ASM P 8742 OUTPUT PR~ SOR 1166 MCP.C P C NOTION CONTROL PROCESSOR 61 MCP_CON.K F - MOTION CONTROL PROOESSOR 61 UMSLNG_4.C P C U.M.S. PROCESSOR 482 CIA PS5.A P 8742 GUlL)kWl~; PROCESSOR(NOT
1 5 S~OWN ) DS5_DEF.A P C DS5000 (S) IN CIRCUIT
EMULATOR

WO 92/0994t ~ O 9 5 4 4 2 PCr/US91 /08892 The following ~rl,yL<ll..../files are used in OUTERLOOP
PROCESSOR 67:
File File Ar, ~ r/
Name ~ Compiler F~n~1~iorl 5 V25TART.ASM P 8086 OPERATING SYSTEM
START--UP
V3_186.ASM P 8086 SPECIAL START-UP CODE
V4 COMT . ASM P 8 0 8 6 SDLC COMMUNI CATION
V41AMX.ASM P 8086 INIT. FOR AMX86 CODE
10 V4_INIT.ASM P 8086 INTERFACE DRIVERS
V4_8 2 7 4 . ASM P 8 o 8 6 SERIAL PORT
DRIVERtl86BD) EK2_DUMMY. C P C INTEL INTERFACING
ER_DUMMY . C P C NORE INTEL INTERFACING
15 RALMAN. C P C KALMAN FILTER
MOVE2WAY . C P C VEHICLE DRIVER
V4ACTPRO. C P C VEHICLE SYSTEM SUBR.
V4CMDT. C P C AGVC COMMAND PROCESSOR
V4DSPMEM. C P C HAND CONTRLLR
DIAGNOSTICS
V4HCOM~qN . K F -- HAND CONTROLLER
CONSTANTS
V4INIT_D.K F -- , INTERFACE DEFINITION
FILE
25 V4LIFT. C P C LIFT SUPPORT
::iUl~KUUllNlS:j V4LIFTT. C P C LIFT TASK CONTROL
V4MESGS . K F -- AGVC MESS . DEF .
CONSTANTS
3 o V4MOTION . K F -- MOTION CONTROL
CONSTANTS
V4MV_WIR. K P C WIRE DRIVER
SUBROUTINES
V4VARS . K F -- MAJOR VEHICLE
- v~RT~T,T~C
V4HCU_AA. C P C HAND CONTROLLER
PROGRAM

WO92/09941 - ~9 4 ~ ~ PCr/l~S9l/08892 V4HCU_CC. C P C ~AND CONTROLLER
PROGRAM
V4MCPEMU. C P C INNERLOOP PROCESSOR
. CONTROL
5 V4MOVT . C P C MOVE TASK CONTROL
V4TMHD.C F C HAND CONTR. SERIAL I/O
V4TRNT . C P C MOTION ~R~NST ~ION
COMMAND INTERPRETER
V4VERS . C P C VEHICLE STATUS
where: P ~c:UL-:6e~1~5 a program ~ile.
F represents a constant or variable ile.
C represents "C" compiler used, otherwise xxxx indicates assembler used.
As seen in the AGV 2A Software Table, outerlooE~
processor 67 comprises a wide variety of ~rUyL~ . Being the master controller, output processor 67 receives and process data from the other mi~ LuuLo~e58uL:~ across bus 816. Each of the software ~r~lyL~h3 and files listed above for output processor 67 performs a specialized function which ranges from initialization of the AGV 2A
mi~:Lu~uLocessor system to calculating complicated KALMAN
error determining and correcting c _~ations involving calibration and correction to AGV i~ LI Lation and motion.
Output processor 67 ~1UYL~Ih~S are assembled using an 8086 assembler or compiled using a C language compiler. Programs V2START.ASM, V3_186.ASM, V4COMT.ASM, V41AMX.ASM, V4_INIT.ASM, AND V4_8274 are assembled using the 8086 assembler. Programs EK2_DUMMY.C, EK_DUMMY.C, KALMAN.C, MOVE2WAY.C, V4ACTPRO.C, V4CMDT.C, V4DSPMEM.C, V4LIFT.C, V4LIFTT.C, V4MV_WIR.C V4HCU_AA.C, V4HCU_CC.C, V4MCPEMU.C, V4MOVT.C, V4TMHD.C, V4TRNT.C, AND V4VERS.C
are compiled using the C language compiler. Special data files provide fixed and variable storage for constants and variables and comprise V4HCOMMN.K, V4INIT D.K, .r, ~ ID~ V~

WQ92/09941 ~ ~ 2 ~ ~ ~ 4 4 2 ~1 PCr/US91/08892 Program V2START.ASM initi~li70~ the total system at powerup and after any manual or automated reinitialization command. Program V3_186.ASM provides 21dditional and uniS~ue startup control for the 80186 5 mi~;Lu~-ùces4uL. Program V41AMX.ASM provides initialization and startup for the AMX86 operating system. Program V4_INIT.ASM comprises a plurality of interface drivers for the microprocessors on digital I/0 board 1194 . File V4INIT_D.K ~U,U,UUL L:i V4_INIT.ASN
10 providing accessible data which comprise assembler code def ining interf ace constants . Program V4ACTPR0 . C
init;~l;70c and monitors AGV 2A in,L- ~ation. In addition, program V4ACTPRO.C i~-ates the setting of visual and audible alarms from output processor 67.
15 Program V4CMDT.C peL~oL. 5 the important task of storing and queuing _ ~ received from AGVC 13. Received '- are parsed for distribution to initiate command-related AGV 2A tasks.
Program V4DSPMEM. C addr =~ses and thereby 20 ;~ccP~o~ memory locations to make available to a display or other user communication device manually selected stored memory data from within a selected mi~;Lu~Lucessor.
As such, program V4DSPNEM.C is used in AGV 2A debugging processes. Program V4HCûM~N.K provides an interface for 25 manual control modules V4HCU_AA.C and V4HCU_CC.C.
Modules V4HCU_AA. C AND V4HCU_ÇC. C each provide direct user control through a manual hand controller to exercise an interconnected AGV 2A as part of manual test ~ruce-lu~ ~5 . Through the use of modules V4HCU_AA. C and 30 V4HCU_CC.C, extensive diagnostics are manually performed on an attached AGV 2A. Program V4TMHD. C supports high level serial ~i cF~tion between the outerloop processor 67 and the hand controller . V4HCûMMN . K
comprises a f ile of constants from which hand control 35 definitions are derived.
AGV 2A lift control support is provided by ~oyL~ V4LIFT.K AND V4LIFT.C. Program V4LIFT.K
-_ _ _ _ _ _ ~ , .. .. _ .. . . _ . . . _ . .. . . .

~WO 92/09941 2 ~ 9 5 4 4 2 Pcr/US9l/08892 interprets manual nAc and provides the drivers for manual control of an AGV 2A lift -- -niF--. Program V4LIFT.C provides automatic controls for the AGV 2A lift ,, ' -ni f---.
Program V4COMT.ASM est~hlichpq protocol for the t half duplex SDLC communications in the AGV 2A. Program V4_8274.ASM performs the driver functions of the SDLC
hardware serial port as required for SDLC communications chip 812. File V4NESGS.R stores definitions of items related to AGV 2A, nAq, _ i ~ations and other codes related to interf ace ~ i cations between an each AGV 2A and AGVC 13.
Programs EK_DUMMY . C AND EX2_DUMMY. C locate the C language _ il Pr generated program modules and where appropriate link the assembled and compiled ~)L O~L ~ . with INTEL library functions.
Program RALMAN. C performs the extensive calculations associated with the Kalman error correcting processes. As earlier described, program RALMAN.C
receives input from a series of redundantly sensed parameters and calculates error correcting inputs to the navigation and guidance in~LL, ~tion. Program RALMAN. C operates in real time providing continuous and timely updates for vehicle guidance.
As part of dynamic AGV 2A control earlier described and provided by outerloop plucessu- 67, program MOVE2WAY.C provides navigation and path control for update marker pathways. In addition, program MOVE2WAY . C
also performs digital guidewire guide path control. In conjunction with MOVE2WAY.C, program V4MV_WIR.K guides the AGV 2A over a guidewire marked route 3. Program V4MOVT. C supervises all AGV 2A motion . Motion interrupts are monitored and related - nAq are parsed by program V4MOVT. C. Program V4MCPEMU. C controls acceleration and deceleration of AGV 2A as well as overall speed control.
Program V4MCPEMU. C also provides velocity and direction of rotation control of fifth wheel 57 and sixth wheel WO 92/09941 2 0 9 5 4 4 2 PCr/US91/08892~

59 . Program V4TRNT . C reprocesses AGV 2A motion ~- n-lc for setup and queuing ~uL~.06es. File V4MOTION.K stores data definitions of parameters used to control and move AGV 2A. File V4VARS . K acces6ibly stores tho6e AGV 2A
vehicle def ining parameters which are common to vehicle motion and command EJ~UyLc~
Program V4VERS . C comprises a table of pa~ Prs which define and identify each vehicle. Such paL ~Prs comprise vehicle version number, vehicle type and other parameters which set the particular AGV 2A
npart from other vehicles used in the AGV 2A control system .
Software for analog input processor 1170, program ANIN.ASM, is written in assembler 8742 language.
In general, program ANIN.ASM samples and store for retrieval each of sixteen analog input line and provides interrupts to outerloop ,ul c,cebcu. 67 based upon predetPrm;nP~l conditions. Specifically, program ANIN.ASM
gatedly samples inputs from each t~rh~ -ter 33, supervises digitization, stores and monitors the samples thereby providing a safety backup to operation of program MCP . C which resides in motion control processor 61.
Further, program ANIN.ASM samples inputs from a joy stick on a manual AGV 2A control box wherefrom an AGV 2A is manually controlled. As part of the monitoring cycle of ANIN.ASM, program ANIN.ASM samples, evaluates inputs from obstacle detectors and the AGV 2A battery voltage.
ANIN.ASM also samples and stores, for retrieval, each of sixteen analog input lines and generates interrupts to outerloop ~LucessuL 67 when predetPrminp~ conditions on the sampled analog input lines occur.
Program CNT.ASM generally processes signals as60ciated with encoder processor 1164. Program CNT.ASM
processes signals from gyro 550, fifth wheel 57 encoder 58 and sixth wheel 59 encoder 58 for further processing by program K~L~AN. C, described above. In the currently preferred ~ -nt, three input rhAnnel s are sampled , . _ . _ . . _ . _ .. ... . .. _ . . . .. , .... . . = _ _ _ _ _ _ _ _ _ W092/09941 2~ ~B44a l : _PCr/US91/08892 for angular and linear travel mea..UL~ . Program CNT.ASM is written and ~c~ 1P~ using in 8742 assembler language .
,. Program DIN.ASM resides in input ~Locei uL 1168 5 and monitor6 forty discrete digital inputs from AGV 2A
borne switches and digital sensors. Program DIN.ASM
~Cc~ccihly 5tores the monitoring derived status bits in groups of eight bit bytes f or later retrieval by outerloop pIucessuL 67. Further DIN.ASM generates an 10 interrupt to outerloop processor 67 each time a monitored input changes state. Interrupts are r~u~ ed upon predet~rminPd configurable conditions to minimi7~
unwanted inputs to outerloop ~ ocesso~ 67 . Monitored digital E;witches and sensors comprise limit switches and g~ y stop apparatus but do not comprise inputs from antennas and update markers . Program DIN . ASM is sPm~led in 8742 assembler language.
Program DOUT . ASM controls and sPq~ nrPc operation of output processor 1168. Generally, program 20 DOUT.ASM controls forty discrete outputs and four pseudo outputs provided for monitoring (no physical AGV 2A
connections are provided to the pseudo outputs). Outputs are discretely set in groups of eight bit bytes. A mask function is used to set, clear, or ignore bits within 2 5 each group being processed by program DOUT . ASM . Program DOUT.ASM comprises timing controls which permit each selected output of the forty-four discrete and pseudo outputs to be ~uy- ' as a pulse, which is preset to be active for a specified time period. Program DOUT.ASM
3 0 conditionally generates an interrupt when a pulse terminates and when predetPrmi nP~l criteria are set and met. In addition, each of the forty-four outputs may be E;elected to be pI uy ~ hly and individually controlled to cycle the selected output at a p~edetermined but 35 individually set frequency. The four pseudo outputs are optionally used as system clocks. Specifically program DOUT.ASM controls operation of devices compri~ing AGV 2A

ao ~944a ~i WO 92/09941 PCr/US91/08892 lights, speaker tone and duration, ~GV 2A beeper audio.
Program DOUT.ASM i8 an 8742 ~sP~hll~r language program.
Program CON.ASM controls SDLC protocol for ; c~tions between each AGV 2A and AGVC 13 . The 5 protocol comprises bu~fering of data to and from the AGV
2A. Program COM.ASM resides in SD1C ; rations mi~uuL~J~es~vI 810 ' and is written and assembled in 8742 lPr language.
Program DSCOMM. C controls serial bus irationS to AGV 2A E~LUCeS~U~. RP-:j11;n~ on serial processor 1192, program DSCOMM. C provides very high speed communications f or each AGV 2A traversing a route 5 marked by update markers 6 and likc paths where high speed serial ; ~ations are used. Program DSCOMM. C
controllably provides a - ; cations interface for all AGV 2A communications except for intl:r ; ratiOn between AGVC 13 and AGV 2A. Program DSCOMM. C receptively acquires data from update marker system processor 482 program ~MSLNG_4.C and systematically transfers the update marker data to outerloop ~Lucessur 67. Program DSCOMM. C is written and - ; l ed in "C" language .
Program UMSLNG_4 . C processes detection and position determination of each LLc.v~sed update marker 6, as earlier described. Acquired information is transferred to program DSCOMM. C which further relays the data to outerloop processor 67 . Program UMSLNG 4 . C is a "C" language program.
Program CIA_PS5.A controls ;cations streams and provides a high speed serial communication interface between AGVC 13 and the floor communications driver of floor controller 13B. Program CIA_PS5.A
resides in off-vehicle processors such a6 a processor located in floor controller 13B. Program CIA_PS5.A is a DS5000 compiled program.
Program DS5_DEF.A provides an assembler interface which facilitates the use of an in circuit emulator (ICE) to test processor memory and hardware.
_ .
..... .... . , . . . . . . _ . . . . _ _ _ . .. . . _ . . _ ...

WO 92/09941 2 0 9 ~ 4 4 2 Pcr/us9l/08892 Program DS5_DEF.A resides in a circuit emulator. Program DS5 HCU. C resides in a DS5000 in the previously mentioned hand control unit and is not listed in the software table as it does not p~r~-n~ntly reside in AGV 2A. Program 5 DS5 HCU.C controls function keys, display output, reading of a keyboard, and i cations to outerloop processor 67 when the hand controller is attached for use. Program DS5 DEF.A is written and compiled in "C" 1 Anq~R~e.
Program MCP. C resides in motion control 10 ~LocessuL 61. Generally, program MCP.C controls analog output to port drive wheel 8 and starboard drive wheel 10 motor controllers . Program MCP . C also provides guidance control when traversing a guidewire route 3 and regulates an emergency stop . In addition, NCP. C monitors the 15 - ,v~ ~ of fifth drive 57 and sixth drive wheel 59 relative to analog output control and generates a fatal stop when an error of a prodotorm; n~c~ magnitude is detected. Program MCP.C ~JL-JyL -hly receives direct but gated inputs via signal paths 1184, 1162, and 1186. File 20 MCP CON.K RCCo~sihly stores code definitions and specif ications used by program MCP . C . Program MCP . c is a "C" language program.
The invention may be ~mhorl;orl in other specific forms without departing from the spirit or essential 25 characteristics thereof. ~he present embodiment is therefore to be considered in all ~ eL LS as illustrative and not restrictive, the scope of the invention being indicated by the RrponAo~l claims rather than by the foregoing description, and all changes which 3 0 come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.
What is claimed and desired to be secured by Letters Patent is:
. .
. ~..
-

Claims (46)

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE
PROPERTY OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. An automatic guided vehicle comprising:
a tri-mode, on-board vehicle navigational and guidance system comprising means for navigating ana guiding over one or more of a plurality of distinct and independent pathway defining means:
the tri-mode navigational and guidance system further comprising:
means for detecting signals from a floor fixed update marker as the automated guided vehicle passes each update marker and means by which the floor location and bearing of the automatic guided vehicle is determined from the update marker signals and exclusively used for guidance without receiving signals from a guidewire;
means for detecting signals from an active guidewire along which the automated guided vehicle is displaced and means by which a bearing to be traversed by the automatic guided vehicle is established from the guidewire signals and exclusively used for guidance;
means for detecting signals from a passive guidewire loop when the automated guided vehicle is disposed within detectable range and means by which the automatic guided vehicle is precisely located at a predetermined site.

2. A method of controlling an automatic guided vehicle comprising the steps of:
a. placing the automatic guided vehicle upon a floor comprising a pathway primarily marked by fixed, signal generating update markers, detecting update marker signals without use of signals from a guidewire as the vehicle moves between update markers, determining floor coordinates defining the location of the automatic guided vehicle from the update marker signals;
b. placing the same automatic guided vehicle upon a floor comprising a pathway primarily marked by an active guidewire from which guidewire signals emanate, detecting the guidewire signals and determining the bearing to be traversed by the automatic guided vehicle across the floor from detected guidewire signals; and c. placing the same automatic guided vehicle upon a floor comprlsing a passive guidewire and a docking site, energizing the passive guidewire by said vehicle and detecting passive guidewire signals at said vehicle and precisely docking the automatic guided vehicle at a docking site using the passive guidewire signals.

3. An automatic guided vehicle system comprising:
at least one floor traversing automatic guided vehicle comprising a top surface permanently affixed atop said vehicle, said surface being free of pathway sensors, and means for detecting signals from floor update markers which define at least in part navigational and guidance information for a vehicle pathway defining active guidewire for another pathway and a vehicle directional sensor comprising tuning fork means;
means for detecting bearing independent of signals from the update markers and the active guidewire, each of said means being disposed below said top surface;
means disposed on the vehicle beneath said top surface and affixed to saia vehicle for guiding and navigating said vehicle based upon signals received from either of said detecting means and said vehicle directional sensor.

4. A system according to claim 3 further comprising Kalman filter means to which a directional output derived from the tuning fork means is communicated to compensate for error due to tuning fork drift.

5. A system according to claim 4 further comprising means communicating a coordinates output to the Kalman filter means to compensate for caordinateg error.

6. A system according to claim 3 wherein the top surface of the vehicle comprises means for loading and lifting.

7. A system according to claim 3 further comprising inertial table means which maintains the tuning fork means in an essentially fixed orientation relative to said floor by which high gain is obtained.

8. A method of providing coordinate and bearing information concerning the location and direction of travel of an automatic guided vehicle traversing a path across a floor, comprising the steps of:
providing a floor-carried coordinate-defining pathway mark over which said vehicle is caused to travel irom time to time;
generating a vehicle tuning fork signal and using the tuning fork signal to acquire a first estimate of the bearing of the vehicle;
processing said tuning fork signal to at least correct for tuning fork drift and thereby more accurately defining the bearing of the vehicle;
detecting a signal emitted by said floor-carried coordinate-defining pathway mark in the immediate vicinity of the vehicle and using said detected signal to ascertain the floor coordinates and bearing of the vehicle .

9. A method according to claim 8 comprising the step of delivering a signal representative of the more accurately defined bearing and a signal representative of the ascertained floor coordinates to an error compensating site.

10. A method according to claim 8 wherein the generating, processing and detecting steps occur in generating and detecting apparatus disposed within and below, but not at the top of the vehicle.

11 An automatic guided vehicle system comprising:
a floor traversed by automatic guided vehicles;
at least one guidewire-controlled first automatic guided vehicle comprising location and bearing-sensing means and guidewire transceiver means;
at least one wireless-controlled second automatic guided vehicle comprising non-guidewire location and bearing sensing means and wireless transceiver means;

floor guidewire means over which at least said first automatic guided vehicle traverses;
control center means comprising control guidewire transcelver means and control wireless transceiver means in communication with both vehicles, in combination the control guidewire transceiver means and control wireless transceiver means comprising means by which control iniformation is communicated, the control information being communicated with compatible data rate transfer characteristics and with identical information content along the floor guidewire means to the guidewire transceiver means of the at least one guidewire controlled first vehicle and wirelessly to the wireless transceiver means of the at least one wireless-controlled second vehicle simultaneously.

12 A system according to claim 11 wherei n the means by which the identical control information is communicated comprise means communicating to both vehicles at substantially the same high baud rate.

13. A system according to claim 11 wherein each vehicle comprises means for sensing and calculating bearing and location of the vehicle and means upon receiving control information for changing the bearing traversed by the vehicle to reach a predetermined floor destination.

14. A system according to claim 11 wherein each said at least one first and second vehicle comprises at least one means for determining bearing and floor coordinates from at least one floor based path defining means.

15. A system according to claim 11 wherein each vehicle comprises means for calculating a path whereby the vehicle is guided to reach a predetermined floor destination at a predetermined bearing from a current location of the vehicle.

16. An automatic guided vehicle system comprising a floor;

a plurality of pathway defining means associated with the floor and comprising at least two of stationary update markers, active guidewlres, and passive guide means each of which is individually and exclusively used to provide guidance signals for the vehicle over at least a portion of the floor;
at least one automatic guided vehicle comprising tri-mode vehicle navigational and guidance means;
the tri-mode navigational and guidance means comprising:
means for detecting the position of the stationary update markers relative to said at least one automatic guided vehicle and determinlng a floor location at which the automatic guided vehicle is disposed from time to time;
means for detecting the bearing of the automatic guided vehicle relative to the floor disposed active guidewire means and determining a bearing of the automatic guided vehicle from time to time;
means for detecting the location and bearing of the passive guide means juxtaposed a docking site and thereby precisely guiding the automatic guided vehicle to a desired location .

17. A system according to claim 16 wherein said passive guide means comprise means removable from said floor and other stationary objects whereby the passive guide means are moved from one place to another to establish a temporary guidepath.

18. A method of controlling an automatic guided vehicle comprising the steps of:
individually and collectively navigating and guiding the automatic vehicles under the direction of a master controller, across a floor comprising at least two of the iollowing steps:
a. navigating and guiding at least one of the vehicles along a pathway defined by stationary, floor-disposed update markers, by determining the floor coordinates at which the one automatic vehicle is located from time to time by detecting the position of successive update markers, storing said coordinates, determining from said determined and stored floor coordinates the bearing and the error in the bearing of the one vehicle and using the error to correct the bearing Df the at least one vehicle;
b . moving at least one of the vehicles along a pathway defired by an active guidewire, by establishing, by guidewire signals, the bearing to be traversed by the one automatic guided vehicle across the floor from time to time; and c. moving at least one of the vehicles along a pathway defined by a passive guide, by precisely positioning the one automatic guided vehicle, at a desired parking site, using localized passive guide signals.

19. A method according to claim 18 wherein the third moving step comprises using power from the at least one automated guided vehicle to provide energy which generates the localized passive guide signals.

20. A method of providing coordinates and bearing information concerning the location of an automatic guided vehicle and causing the vehicle to traverse a path across a floor to a target destination comprising the steps of:
providing within the vehicle a means for calculating a guidepath, based upon the target destination and current floor coordinates and bearing of the vehicle, along which a guidance and navigation system contained within the vehicle causes the vehicle to traverse;
sending the target destination comprising target floor coordinates and a target bearing from an off-vehicle control center to the vehicle;
generating a tuning fork signal at the vehicle defining the bearing of the vehicle, processing the tuning fork eignal to correct for drift and delivering the corrected tuning fork signal to the guidance and navigation system;
sensing a floor coordinates signal at the floor in the immediate vicinity of the vehicle, determining the position of the vehicle relative to the floor coordinates and delivering the sensed floor coordinates to the guidance and navigation system;
calculating the guidepath along which the guidance and navigation system causes the vehicle to traverse from the target destination, the position of the vehicle and bearing from the corrected tuning fork signal.

21. A method according to claim 20 further comprising the step of processi}lg and bearing signal in combination with past and current floor coordinates signals to compensate for vehicle location and bearing errors.

22. A method according to claim 20 wherein the generating and creating steps occur in apparatus disposed below and within, and not at the top surface of the vehicle.

23. A method according to claim 20 further comprising the step of maximizing gain and minimizing error during said generating step by controlling the directional orientation of the tuning fork from which the tuning fork signal is emitted.

24 . A method of communicating and controlling vehicle travel in an unmanned automatic guided vehicle system comprising the steps of:
providing a guidewire pathway disposed in a floor over which at least one guidewire controlled first automatic guided vehicle travels and a non-guidewire marked pathway in the floor over which at least one second automated guided vehicle controlled by wireless communications travels;
selectively issuing compatible guidewire and wireless communications control signals comprising identical control information transmitted at a single data rate from a control center to the at least one guidewire controlled first automatic guided vehicle and to the at least one wireless controlled second automatic guided vehicle, respectively;
controlling travel of the first vehicle using communications control signals received by the first vehicle from the control center through the guidewire at the first vehicle;
controlling travel of the second vehicle exclusive of the guidewire by using communications control signals received by the second vehicle through wireless communications and sensing location and bearing of the second vehicle using non-guidewire guidepath marker navigational information derived at the second vehicle.

25. A method according to claim 24 further comprising the steps of using active guidewire signals to correct the bearing of the first vehicle to cause the first vehicle to reach a desired destination.

26. A method according to claim 24 further comprising the steps of correcting the bearing of the second vehicle at the second vehicle to cause the second vehicle to reach a desired destination.

27. A method according to claim 24 further comprising the steps of determining the error in any estimate of locations of the first vehicle at the first vehicle and congruently correcting the displacement of the first vehicle under control of a navigation and guidance system disposed at the first vehicle.

28. A method according to claim 24 further comprising the steps of determining the error in any estimate of location of the second vehicle at the second vehicle and congruently correcting the displacement of the second vehicle under control of a navigation and guidance system disposed at the second vehicle.

29. A method according to claim 24 wherein the second controlling step comprises sensing at the second vehicle the bearing of the second vehicle via an angular rate tuning fork sensor disposed on the second vehicle.

30. A method according to claim 29 wherein during a first period, a third unmanned automatic guided vehicle performs as the first vehicle in the first travel controlling step and, during a second period, the third vehicle functions as the second vehicle in the second travel controlling step.

31. A navigation and guidance system for an automated guided vehicle comprising:
means disposed on said vehicle for receiving at any point along a vehicle rcute a target waypoint position and vehicle exit bearing from said waypoint from an off-vehicle controller;
means disposed on said vehicle or selecting and calculating a guidepath to be transversed by said vehicle from a plurality of possible guidepaths based upon a then current vehicle position which comprises an origin for and bearing of the selected guidepath and the target waypoint position and vehicle exit bearing from the target waypoint;
means for constraining said vehicle to the selected guidepath such that the origin of the selected guidepath comprises the existing vehicle bearing and location and the target waypoint position and said vehicle exit bearing away from the target waypoint .

32. A method of causing an automatic guided vehicle to reach a predetermined destination comprising the steps of:
(a) providing a floor comprising a pathway marked by at least one update marker for non-guidewire navigation and guidance of the vehicle, said pathway comprisina the predetermined destination;
(b) sending wireless control signals from a central control system to selectively direct the automatic guided vehicle along at least a portion of the pathway toward the predetermined destination;
(c) receiving and utilizing wireless control signals at the automated guided vehicle;
(d) optionally obtaining floor related, vehicle generated coordinated signals through vehicle detection of information from the update marker;
(e) communicating the vehicle-generated coordinate signals and the controlled signals to a vehicle-borne navigation and guidance system;
(f ) calculating, to the navigation and guidance system, a guide path comprising exit target coordinates and an exit bearing for tke vehicle along said portion of the pathway along which the vehicle is to be navigated and guided; and (g) repeating steps (b) through (f) until the vehicle reaches the predetermined destination.

33. A method according to claim 32 wherein step (c) comprises the additional step of receiving and utilising guidewire communication signals in causing the vehicle to reach a predetermined location.

34. A method according to claim 32 wherein the obtaining step (d) further comprises deriving bearing related signals from a vehicle tuning fork gyro for use by the navigation and guidance system.

35. A method according to claim 34 wherein step (d) further comprises the step of controlling the orientation of the tuning fork gyro using an inertial platform.

36. A method according to claim 32 wherein the obtaining step (d) comprises deriving signals from wheel encoders which measure differential in wheel rotations for use by the navigation and guidance system.

37. A method according to claim 32 wherein the obtaining etep (d) comprises obtaining signals from stationary update markers along one portion of the pathway and obtaining signals from a guidewire along another portion of the pathway.

38. An automatic guided vehicle system for an autonomously-operating automatic guided vehicle, said system comprising:
means disposed in the autonomously operating automated guided vehicle for nulling navigation and guidance bearing and position errors, said means comprising:
at least two measuring means for redundantly estimating bearing of the vehicle, and for redundantly estimating a position of the vehicle;

at least one of said measuring means comprising means for detecting an update marker disposed in a floor below said vehicle;
means for receiving said estimates of bearing and position from said measuring means and assimilating redundantly measured position and bearing values:
Kalman filter means for processing the redundantly measured position and bearing values to provide error nulling corrections comprising at least one state variable;
means for navigating and guiding the automatic guided vehicle across the floor based upon error-nulling corrections provided by the Kalman filter means.

39. A method for controlling movement of an unmanned vehicle which operates in a self-contained guidance mode whereby the vehicle operates up to a given speed and moves from one path segment to another path segment without slowing down, comprising the following steps:
(a) sending from a non-vehicle source at least one move command comprising at least one vehicle target destination and direction for the unmanned vehicle sent from a non-vehicle source;
(b) receiving the at least one more command by the unmanned vehicle;
(c) decoding in a background area of a multi-making operation in a computer on-board the unmanned vehicle, the at least one move command related to the vehicle target destination and direction of the unmanned vehicle;
(d) transferring the at least one decoded move command to an on-board vehicle controller to cause the unmanned vehicle to move toward a predetermined target destinator;
(e) repeating steps (a) through (d) whereby each next move command is received timely such that the vehicle continues without slowing down.

40. A method according to claim 39 wherein the given speed is on the order of 200 feet/second.

41. A method for initially acquiring a signal from a guidewire by a hybrid unmanned vehicle which operates by self-contained navigation and guidance with occasianal calibrating updates or, alternatively, by following a guidewire, comprising the following steps:
acquiring a signal emitted by the guidewire when the guidewire is first disposed below the vehicle in a detectable range of guidewire signal detecting antennas on the vehicle, but while the position of the guidewire is at an actual offset position relative to a longitudinal center line of the vehicle;
defining the actual offset position upon guidewire signal acquisition to be a desired relative center line-to-guidewire offset position; and over a period of time, making small incremental reductions in the magnitude of the desired relative vehicle center line-to-guidewire offset position, thereby driving the vehicle to a position relative to the guidewire where the actual offset is zero.

42. An automatic guided vehicle system comprising:
at least one pathway marked exclusively by one type of pathway defining means derived from a group of different types of pathway-defining means, the different types of pathway-defining meaffs comprising floor disposed update markers defining one type of pathway and active guidewires defining another type of pathway;
an unmanned automatic guided vehicle comprising on-board automated guided vehicle navigational and guidance means comprising means for navigating and guiding over each of said types of pathway defining means;
the navigational and guidance means further comprising:
means for detecting individual update markers as the automated guided vehicle passes across each update marker and by which the floor location and bearing of the automatic guided vehicle is détermined to the exclusion of guidewires from time to time;
means for detecting active guidewires as the automated guided vehicle is moving along a pathway marked by active guidewires and by which a bearing to be traversed by the automatic guided vehicle is established.

43. A method of controlling at least one automatic guided vehicle comprising the steps of:
providing at least one pathway marked exclusively by a type of pathway defining means derived from a group of different types of pathway-defining means comprising floor disposed update markers and active guidewires;
providing on-vehicle automated guided vehicle navigation and guidance means for navigating and guiding over each of said pathway-defining means exclusive of other types of pathway-defining means;
navigating the at least one automatic vehicle comprising the navigation and guidance means across a floor by:
a. during a first period while the at least one vehicle is moving along a pathway marked exclusively by update markers, detecting an update marker as the vehicle moves within detectable distance of the marker and thereby determining the floor coordinates of the at least one automatic guided vehicle from time to time;
b. during a second period while the at least one automatic guided vehicle is moving along a pathway marked by an active guidewire, detecting guidewire signals from the active guidewire. and thereby establishing the bearing to be traversed by the automatic guided vehicle across the floor.

44. An unmanned automatic guided vehicle system comprising:
at least one pathway marked exclusively by a type of pathway-defining means derived from a group of different types of pathway defining means comprising floor disposed update markers and passive guidewire loops;
at least one unmanned automatic guided vehicle comprising on-board automated guided vehicle navigational and guidance means comprising means to the exclusion of other types of pathway-defining means;
the navigational and guidance means further comprising:

means for detecting each of said update markers as the automated guided vehicle passes across each update marker and by which the floor location of the automatic guided vehicle is determined from time to time;
means for energizing and detecting a passive guidewire loop when the automated guided vehicle is disposed within energizing and detecting range of a pathway marked by the passive guidewire loop and by which the automatic guided vehicle is precisely located at a predetermined site.

45. A method of controlling at least one automatic guided vehicle comprising the steps of:
providillg at least one pathway marked exclusively by a type of pathway-defining means derived irom a group of different types of pathway-defining means comprising floor disposed update markers and passive guidewires;
providing on-vehicle automated guided vehicle navigation and guidance means for navigating and guiding said vehicle over each of said pathway defining means to the exclusion of other types of pathway-defining means;
navigating and guiding an automatic vehicle across a floor by:
a. during a first period while the at least one vehicle is moving along a pathway marked by update markers, detecting an update marker as the vehicle moves within detectable distance of the mark and thereby determining the floor coordinates of the at least one automatic guided vehicle from time to time;
b. during a second period, while the at least one automatic guided vehicle is moving along a pathway at a docking site marked by a passive guidewire, energizing a passive guidewire by said vehicle and detecting a signal from said energized passive guidewire by said vehicle and thereby precisely docking the automatic guided vehicle at the docking site using the localized passive guidewire signal.

46. An unmanned automatic guided vehicle system comprising:

at least one pathway exclusively marked by one type of pathway defining means derived from a group of different types of pathway-defining means comprising floor disposed update markers, active guidewires and passive guidewire loops;
at least one unmanned automatic guided vehicle comprising on-board automated guided vehicle navigational and guidance means comprising means for navigating and guiding over a plurality of said pathway-definlng means.