CA2114648A1 - Navigation and positioning system and method using uncoordinated beacon signals - Google Patents

Abstract

A positioning system uses a multiplicity of commercial broadcast stereo FM radio signal transmitters (102), at known fixed locations, each of which transmits a beacon signal having a phase that is unsynchronized with the phases of the beacon signals of the other transmitters. All the beacon signals have a frequency approximately equal to a 19 KHz. A fixed position observer unit (110), positioned at a known location, receives the beacon signals from all the transmitters in the vicinity, determines their relative phases, and broadcasts data representing these relative phases.
Mobile units (120), at unknown locations, receive these broadcast values, as well as beacon signals from at least three radio transmitters. Each mobile unit includes phase detection circuitry (144) that detects the phases of the beacon signals at the mobile receiver's current position. This is accomplished using a signal radio signal receiver. A digital phase-locked loop, coupled to the radio signal receiver, generates a phase error signal for each beacon signal. The phase error signals are then used to compute a distinct phase value for each beacon signal. In the preferred embodiment, each mobile receiver includes a computer (150) for computing its location based on the broadcast relative phase values and the detected phases. In another embodiment the position computations for many mobile receivers are performed at a central data processing station.

Description

WO 93/04378 PCl/US92/068~4 i .
6 ~ 8 NAVIG~TION AND POSITIONING SYSTEM AND METHOt) USING UNCOORDINATED BEACON SIGNALS

The present inven~ion relates generally to navigation and positioning systems in which an object or user at an un'~no~wn location receives broadcast signals ~rom several sources and uses information derived there~rom to determir)e the object's or user's current position. More particularly, the present invention relates 5 to a positioning system which uses uncoordinated beacon signals ~rom commercial radio broadcasts to enable a highly accurate position determination.

BACKGROUND OF THE INVENTION

10 Most of the well known prior art navigation and positioning system, such as LORAN, NAVSTA~, and GPS use special transmitters, either orbiting the earth or land-based, dedicated solely to ~he positioning system. Great efforts are made in such systems to synchronize ~he transmit~ers in these systems so tha~ the phases of the beacon signals reaching a user at any location on the earth's 1~ sur~ace can be calcula~ed with a known level of accuracy. These ~ypes o~ prior art systems generally require huge capital investments, often government subsidized, because the transmitters and their control systems are very complex and expensive.

w0 93/04378 i~ 6 4 8 PCr/US92/06824 In the present invention, the signals used for positioning are tr,~ ~P KHz pilotsignals transmitted by commercial broadcast stereo FM stations. In the United States, there are nearly 5000 commercial broadcast stereo FM radio stations, and the FCC requires that all commercial stereo FM radio stations broadcast 5 a 19 KHz pilot sub-carrier signal. The maximum allowed deviation is ~ hertz.
Generally the rate of frequency drift has been found to be much less than 0.1 hertz per day. In any case, the transmitters used by the present invention are free, which makes the positioning system of the present invention relatively inexpensive to implement.
Two U.S. patents which describe ~delta-position~ systems using commercial broadcast transmissions are U.S. Patent Nos. 4,054,880 (Dalabakis et al.) and 3,889,264 (Fletcher). These prior art patents describe systems using three space~apart radio signals, each of which is an independent AM radio signal.
15 The systems typically have a vehicle carried mobile receiver, with a separatetuner for each station, and a second receiver at a fixed, known position.
Basically, these systems count ~zero crossings~, each of which indicates that the user has moved a certain distance from his previous location. Thus, if the user needs to know his current position, the user needs to first specify his starting 20 position. A ~ixed position receiver detects ~requency drift of the ~ransmitters, and that drift information is used to adjust and coordinate the zero crossing countsmade by the mobile receivers.

These are ~delta-position~ systems because they determine only the distance 25 and direction travelled by a mobile user ~rom any particular starting point. Neither Dalabakis nor Fletcher actually determines the position of the mobile user, and in fact such a determination is not possible using the Dalabakis or Fletcher systems because these systems do not have the ability to determine the phases of the transmrtted radio signals. Furthermore, since only zero crossings are 30 counted, the positioning accomplished by these systems have ~granularity~, which w0 93/04378 ;~ 8 PCI/US92/06824 in the case of the systems disciosed in these two patents is on the order of thl.~y meters.

Like Dalabakis and Fletcher, the present invention uses spaced apart, commercial5 radio signals. Ths present invention also uses a ~ixed position receiver to help the mobile units determine their position. However, the present invention, unlike Dalabakis and Fletcher, determines the user's position without need tor any starting point information, and determines such positions with a high degree of accuracy. To do this, the fixed position receiver not only determines frequency 10 dritt, it also determines the relative phases of the various beacons with a very high degree o~ accuracy (e.g., within about 0.02 degrees, or equivalently, within about 0.00035 radians) using a digital phase-locked loop. Using this relative phase information, the commercial radio signal beacons are transtorrned, in essence, into coordinated beacon signals with well detined phase relationships.
15 As a resul~, the position o~ the mobile users can be computed from the radio signals received by the mobi!e user, with an accuracy of about ~3 feet in any direction.

Two other important differences between the present invention and Dalabakis 20 and Fletcher are (1) the use of 19 KHz beacon su~carrier signals o~ commercial broadcast stereo FM stations, and (2) the use of a single receiver for observingthe phases of multiple radio signals. FM radio signals are inherently better than the AM radio signals, because FM modulated signals all less susceptible to noise.
Further, the hard limiter found at the front of most FM receivers rejects weaker25 signals, thus providing multipath rejection.

Also important is the use o~ a single receiver which scans through all the available FM radio station frequencies. This is not simply an economy measure. The key here is that precisely measuring the relative phases o~ several signals 30 requires that the measurement system not introduce errors into the system - such i 6 ~ & ~

as the errors that would be caused by having different propagation delays for separate, muHiple receiving circuits. A small phase error can produce a relatively large positioning error. By using the same physical circuitry ~or ali signals, the same phase delay through the circuitry occurs for all received ~ignals.

SUMMARY OF THE INVENTION

In summary, the present invention is a positioning system, which uses a multiplicity of commercial radio signal transmitters, at known fixed locations, each 10 of which transmits a beacon signal having a phase that is unsynchronized withthe phases o~ the beacon signals o~ the other transmitters. All of the beacon signals have a frequency which is approximately equal to a predetermined target trequency. In the pre~erred embodiment, the beacon signals are 19 KHz pilot tones generated by commercial broadcast stereo FM stations. -A ~irst receiver, known as the ~ixed position observer, is positioned at a knownlocation. The fixed observer receives ~he beacon signals, determines the relative phases of the beacon signals, and broadcasts data representing these relative phases. Mobile receivers, at unknown locations, receive these broadcast values, 20 as well as beacon signals ~rom at least three radio transm~ers.

Each mobile receiver includes phase measuremen~ circuitry that detects the phases of the bPacon signals at the mobile receiver's current position. This Is accomplished using a single radio receiver Sor receiving beacon signals at multiple 2~ distinct carrier frequencies. In the pre~erred embodiment, the phase measurement circuitry is implemented with a dig~al phas~locked loop (DPLL), including a digital phase detector coupled to the radio receiver and a loop filter implemented in software. The dig~al phase detector generates a phase error signal for each beacon signal. The DPLL is agile in the sense that a phase measuremen~ for ~0 each beacon signal is generated in a time sequential manner.

WO 93/04378 PCI'/US92/06824 6 ~ 8 ~

In the preterred embodiment, each mobile receiver computes its location based on (A) the relative phase and frequency data broadcast by the fixed position observer, (B) the phases detected by the mobile receiver, and (C) the known positions of the transmitters and the tixed observer. In other embodiments, the phase information generated by the mobile receiver could be transmitted to a computer at a remote location, so that the position computations for many mobilereceivers could be performed at a central data processing station. The position computations may occur in a timely manner, orthey may be deferred and stored.

10 Even though the present invention determines the mobile receiver's absolute position with respect to a detined coordinate system, the invention does not require synchronization of clocks or tifi~e values between the mobile receivers and the fixed observer or any other time base. Rather, time is treated by the mobile receiver as an independent variable, much like the mobile unit's spatial 15 coordinates (e.g., x, y, and z). Usingthe rhethod otthe present invention, relative time offsets between the fixed observer and the mobile receiver, as well as the mobile receiver's absolute position, are determined and maintained by the mobilereceiver at a high level of accuracy.

BRIEF DESCRIPTION OF THE DRAWINGS

Additional objects and features of the invention will be more readily apparent from the following detailed description and appendéd claims when taken in conjunction with the drawings, in which:
Figure 1 schematically depicts a positioning system in accordance with the present invention.

Figure 2 is a block diagram of a mobile receiver unit.

WO g3/04378 . P~/US92/06824 2 ~

Figure 3 is a flowchart of a the main control routine of a mobile receiver unit.
Figure 4 is a block diagram of a digital phase detector for use in both mobile receiver units and a fixed position observer unit.
S
Figure 5 is a circuit diagram of a power estimator circuit.

Figure 6 is a circuit diagram of a numerically controller oscillator (NCO).

10 Figure 7 is a circuit diagram of a phase error accumulator circuit.

Figure 8 is a block diagram of a loop filter routine executed by the CPU in mobiie receiver units, and also by the CPU in a fixed position observer unit.

15 Figure 9 is a block diagram of a loop filter subroutine.

Figure 10 is a block diagram of the phase master accumulator software routine.

Figure 11 is a block diagram of a low pass filter routine.
Figure 12 is a flow chart of the main routine ~or the fixed observer unit.

DESCRIPTION OF THE PREFERRED EMBODIMENT -Referring to Figure 1, the positioning system 100 uses 19 kiloher~z (KHz) pilot signals transmitted by existing commercial FM radio stations 10~. Every commercial broadcast stereo FM station in the United States has a pilot signal, herein called a beacon, that runs at a rate of 19 kilohertz, plus or minus a frequency deviation of no more than two Hertz. There is no coordination of the phases of these signals between stations. FM radio station towers tend to be WO 93/04378 ~ 6 1 8 PCr/US92/06824 spaced around the periphery of major metropolitan areas, creating a set of transmitted beacons suitable ~or dense metropoldan use.

A fixed position observer 1 10 at a known position sequentially scans the FM radio band, receiving all the FM station pilot signals in the vicinity. Since its position `
is known, it can determine the relative phases of all the pilot signals, in terms ot their phases at the point of broadcast. It also computes the frequency drift ot each station, which is the difference between the station's pilot signal frequency and its nominal frequency, 19 KHz. Periodically (e.g., once every 0.5 seconds) it broadcasts reference data representing the relative phases and drift rate values o~ all the FM pilot signals in its vicinity.
:.
In the preferred embodiment, the fixed observer broadcasts model coefficients which are used bythe receiving mobile unit in a shared predictive model of each beacon's drift. The use ot these model coefficients reduces the rate at which reterence data values need to be transmitted while still maintaining a high level of accuracy. This is particularly useful in applications in which the data link between the fixed observer and the mobile unit may be subject to periodic dropouts or is capacity constrained.
In most of the world, the base band of each commercial broadcast stereo FM
station has a first frequency region in which the stationts main program is transmitted, and a subsidiary frequency region, called the ~ubsidiary Communication Authorization (SCA) channel, in which additiorlal programming can be transm~ted. In the preferred embodiment, the fixed position observer 1 10 broadcasts reference data using the SCA channel of at least one FM station,and for greater reliability the SCA channels of at least two FM statisns in eachregion th~t the system 100 is to be used. Alternately, the fixed posdion observer 110 could have its own radio transmitter for broadcasting re~erence data, or it could ùse any other available communication mechanism.

8 ~""'t~

The system 100 can support an unlimited number ot mobile receiver units 120 without increasing the amount of reference data required to be transmitted from the fixed observer unit 110. Each mobile unit 120 receives the broadcast reference data, as well as pilot signals from at least three (and preferably four 5 or more) stations. By using the broadcast reference data to mathematically adjust the received FM pilot signals, the FM pilot signals are, in effect, coordinated.As a result, each mobile unit can compute its position and time.

Referring to Figure 2, each mobile receiver unit 120 has an antenna 140 and 10 a tuner 142 for receiving selected FM radio station signals and extracting its 19 KHz pilot signal trom the baseband signal. A digital phase detector (PD) circuit 144 receives this baseband pilot signal. It also receives clock signals from a clock generator 146, as well as a starting phase value ~ and a ~delta phase~
value ~ ~rom the unit's central processing unit (CPU) 150. These clock, phase 15 and delta phase values are used by the PD 144 to generate ~ reference signal.The PD 144 compares ~he received baseband pilot signal with the reference signal, and generates corresponding phase error information.

In the preferred embodiment, the phase measurement circuitry is implemented 20 with a digital phase-locked ioop (DPLL), which includes the PC) 144 working in conjunction with a loop filter implemented in software (i.e., running on CPU 150).

Tuner 142 is also used to receive broadcast reference data from a selected FM
radio station's SCA channel, and to send the received phase data to CPU 150.
2~ The tuner 142 is by no means constrained to receiving only the re~erence data~rom one or more active SCA channels. It could also be used to receive other transmitted information of use to the mobile unit, or the system in which the mobile unit is used. An example of such use would be periodically updating the data base of antenna locations kept by the mobile unit. Alternately, a second WO 93/04~78 PCI/US92/06~24 ;
~ 6 ~ 8 ., ~ln~r of similar design to tuner 142 could be employed solely ~or the reception o~ reference data.

The CPU 150, under software control, uses phase error information received 5 from the PD 144 to compute and store the phase ~P (at the mobile unit's current position) and delta phase ~ of each beacon signal. The software routines 162-168 and data tables 170 used by the CPU to control operation of the mobile unit are stored in a data storage device or devices 180. For example, the software routines may be stored in read only memory, while data tables and other10 variables are typically stored in high speed random access memory. In the preferred embodiment, the CPU 150 is a 68000 microprocessor made by Motorola, although the initial prototypes~f the invention were implemented usingthe 68030 microprocessor.

The mobile unit 120 may include a ~isplay 182 and keyboard 184, or equivalent user interface. In a typical application of the present invention, the user's position will be displayed as a position on a map, wi~th the user being given the abilityto zoom the displayed map in and out so as to be able to get different perspectives on the user's location. Altemately, in another application the mobile Z0 unlt's present or past position(s) may be displayed at a remote location or time on one or more devices. Such displays may be automatically updated periodically, or as requested by a system user. The mobile unit's position may be transmitted in response to requests, or when certain events occur (such as the mobile unit 120 arriving at a selected location).
2~
MAIN CONTROL ROUTINE - MOBILE UNIT

Re~erring to Figure 3, the main control routine 160 of the mobile unit operates as ~ollows First, the mobile unit t20 tunes to the SCA channel on which the 30 fi%sd observer unit 110 is transmitting its data, and updates its transmitter data WO 93/04378 PCr/US92/06824 - ~
2~1 16~8 `~
10- :::
table 1, ~ ~vith the received reference data values (step 200). The reference data ~rom the fixed observer unit is an interrnittently transmitted stream of data values in a form such as:
cbeacon id 1, observed phase and freq. values 1, ret. time 1~ ;
cbeacon id 2, observed phase and freq. values 2, ref. time 2>
~beacon id 3, obsenJed phase and freq. values 3, ref. time 3>
... .

This data is herein referred to as the fixed obsenJer reference data, or simply 10 as the reference data. The reference data parameters represent the coefficients of a linear model of each beacon's phase, as a function of time t as defined by the fixed observer. The invention is not limited to a linear model, and those skilled in the art will recognize that other modeling is possible The generationof linear model coefficients from the phase values measured by the fixed observer 1~ is discussed later in the section entitled ~Fixed Position Observer Unit~.

The parameters of the reference data relate to the linear model of the beacon phase, which can be wr~en in point-slope form. The linear mooel is:

~n(tu) = ~n-(tu-to) - ~n(t0) (Eq. 1) where n represents the n~h beacon t is the reference time as defined by the fixed observer 2~ tu is t-~, where ~ is an offs~t due to receiver propagation delays and clock bias.
to is an arbitrarily chosen reference time ~3n(tu) is the estimated phase of the nth beacon signal at its antennae as predicted by the fixed observer Rn is the distance of the fixed observer from the nlh beacon's antenna.

wo g3/04378 ~ PCI`/US92/06824 . .

is the wavelength of the nlh beacon signal (approximately 15,789.5 metsrs) (to,~n(to)) is the model's point parameter at an arbitrary instance to of th~ ~ixed observer time t.
~n(tO) is the phase at the nth beacon's antenna, which has been adjusted by the fixed observer by removing ~rom the beacon's measured phase the value 21~R"/~, where ~,, is the wavelength of the n~h beacon.
~I)n iS the model's slope parameter, which is equal to the radian trequency of the n~h beacon signal.

It can be seen that by using this model, the mobile observer can ~erive an estimate of the n'h beacon's phase at an arbitrary time t.

15 The reference data need not be limited to just the model coefficients. Those skilled in the art will recognize that other data for ancillary purposes could also bP included. in the préferred embodiment, the trans~er or communication of the reference data is done in such a manner as to minimize the amount of bandwidth required.
The structure of the beacon data table stored in the mobile unit's memory 180 is shown in Table 1.

Note that in Table 1, Beacon_lD is the frequency, in MegaHertz, of the carrier 2~ signal. T_LOCTN indicates the position of the beacon's transmi~ter in terms of a predefined coordinate system. X_Phase is the phase o~ the signal at the transmitter, at time To~ which is a specific instance of time t, as de~ined by the fi~ed observ~r unit. Drift is the difference between the frequency o~ the beaconsignal and the mobile unit's 19 KHz local clock. C_Phase is the phase of the 30 beacon at the mobile unit's current position at the time of the last 100 millisecond w0 93/04378 ~ 4 ~ Pcr/US~2/o682 time increment. Ac_Phase is t";~ total accumulated phase of ths beacon signal, a~er subtracting off the mobile unit's 19 KHz local clock.

Beacon ID T LOCATN X Phase ~ _ Drift C Phase Ac Phase ...
096.5 Xl,Yl,~l 5.11234 239.1 0.988 1.0234g 43.12224 ...
102.1 X2,Y~,Z2 1.23339 239.2 0.010 4.49876 54.00987 ...
... ... ... ... ... ... ... ...

Next, at step 202, the mobile unit scans FM stations to deterrnine the quality o~ the beacons of interest. The beacons o~ interest can be derived trom several sources, including A) all possible FM ~adio frequencies, B) those listed in the 15 beacon data table 170, and C) those contained within the data received from the fixed observer unit 110. The quality of the beacon is comprised o~ two separate quality considerations of merit: (A) the c~liber of the phase measurements made from the station's pilot tone, and (B) the contribution of error in the navigation solution due to the physical location of the mobile observer 20 with respscttothe beacon antennae.

The caliber of the phase measurements are d~rived from me~surements o~ the phase error signal's variance. A iarge varianc~ would indicate a poor caliber station, whereas a small variance wouid indicate a high caliber station. The 2~ caliber of the phase measur~ments is also influenced by how well ~he DPLI is tracking the beacon. The DPLL will require a finite time period to acquire a new beacon. During this time, ~he caliber of the phase measurement is less than after acquisition. Additionally, acceleration of ~he rnobile observer will cause a fini~e tracking error wlthin the phase-locked loop. During periods of acceleration, 30 the caiiber is less than during periods o~ constan~ velocity.

WO 93/04378 i~ 6 '1 8 PCI`/US92/06824 The contribution of error in the navigatio.. ^nlution due to the physical locations of the beacons' antennae is termed geometrical dilution of precision (GDOP).
GDOP results ~rom the magnification of normally inconsequential errors.
Generally, GDOP will be minimized by a selection of beacons that are positioned 5 at the widest possible spread of directions from the mobile ùnit. Note that if the mobile unit's position has been previously computed in the recent past (typically within the last couple ot seconds), the approximate position of the mobile unit is knom, even if the mobile unit is moving, with much more accuracy than is needed tor selecting a set ot beacons.
If the mobile unit's position is being computed for the first time since the system was powered on or reset, beacons are-selected simply on geographic spread;
a better set ot beacons might be selected, after the mobile unit's position has been computed, on the next pass through the main routine.
If less ~han three of the stations have acceptable signa, quality (step 204), anerror message is sent to the user interface (step 212), and the main routine restarts at its beginning. Othe~ise, the stations used for navigation computations are those with the highest quality.
Even though a minimum of three selected beacons will be used at any one time to compute the mobile unit's position, it is pre~erred to monitor additional spare stations so that if reception of one of used stations deteriorates, the mobile unit is able to immediately switch to a 'spare~ station for the purpose of computing 2~ its position. The phases of the spare stations are monitored along with the phases o f the stations being used to compute the navigation solution Additionally, the spare stations can be included in the navigation computation to improve the accuracy of its solution. The multiple beacon tracking ability o~the [)PLL provides the spare station capability at no additional cost.

W0 93/W378 2 1 1 4 ~ 4 8 PCI`/US92/06824 -14- :
Next, the tuner is sequentially tuned to each o~ the selecl cd ~acons for a period -~
of 100 milliseconds per station. Considering the handling of a particular beaconduring this 100 millisecond period, the first 50 milliseconds are ~tilized to allow the tuner to settle, and during the second 50 milliseconds, phase error values flrom the PD 144 are accumulated in a register 409 (see Figure 8) for use by the loop filter routine 406 (see Figures 8 and 9). The accumulated phase error values are then used to compute the drift frequency, the delta phase, and the phase of the pilot signal (step 216). These phase values then update that particular beacon's data in the beacon data table 170. Also computed and stored in the beacon data table 170 are its current phase and accumulated phase (C_Phase and Ac_Phase in Table 1).

The other selected beacons are then adjusted using their previously collected data v~lues so that their accumulated phase values all represent the phase of the corresponding pilot signal at consistent instan~es in time. This is done simply by adding or subtracting the phase shift associated with the proper number of 100 millisecond inten~als. Note that if the pilot signal's frequency were to be exactly 19,000 herk, the pilot signal's accumulated phase at time T and its value -at time T ~ 100 milliseconds arethe same. Therefore, to adjust a beacon signal'sstored phase value for the passage of 100 milliseconds of tim~, the mobile unit's CPU just adds 0.1 times the value of the pilot signal's drift frequency to the previously stored phase value. ;

The DPLL is made agile by the combination of tuning the radio receiver and 25 loading the corresponding beacon's current phase value (from the beacon data table 170) into a digital numerically controller oscillator (NCO) 260 (Figure 4).

The manner in which the frequency and phase of each pilot signal are computed is discussed in more detail below with respect to Figures 4 through 11.

`- 211 16 i8 - 1s- ..
On a periodic basis (e.g., once per second), a navigation routine 168 is ex~cuted that computes the mobile unit's current position (step 218) based on the data stored in the beacon data table 170. The measurement of phase by the DPLL
continues to operate in parallel with the navigation routine.
PHASE DETECTOR CIRCUIT

The digital phase-locked loop (DPLL) used by the preferred embodiment is implemented with a digital phase detector 144, and a software loop filter. The DPLL can measure a beacon's phase with an error on the order of 0.000096 radians. Since the wavelength of a 1~KHz signal is about ten miles (actually 15,789.5 meters), this level of precision is necessary to achieve a resolution of about one ~ourth of a meter.

1~ A sin~le phase detector 144, coupled to a single receiver/tuner, is used to process all the selected pilot signals. 8y using the same physical circuitry for all signals, ;~
appro%imatelythe same phase delaythrough the circuitry occurs for all received signals. This common mode phase delay is eliminated by the navigation algorithm. A differential mode phase error of even 0.01 radians, due to differing propagation delays for distinct tuner and/or phase detector circuits, would cause a positioning error on the order of 2~.1 meters. Thus, using a single tuner and phase detector helps to minimize positioning error.

Referring to Figure 4, inputs to the phase detector (PD) 144 are the baseband pilot signal from the tuner 142, plus the following signals:
CLK a clock signal running at 152 KHz (i.e., eight times faster than a 19 KHz pilot signal);
MTM a clock signal running at 1 KHz and which is enabled during only one of each 152 cycles of ~he CLK signal;
MTM the inverse of MTM;

WO93~04378 ~ PCI/US92/06824 the last computed delta phase o~ the pilot signal being processe~J
and ~P the predicted phase o~ the pilot signal at the be~inning of the next one millisecond signal processing period.

The clock generator 146 uses an oscillator 250 and divider 2~2 to generate the 152 KHz CLK signal, followed by a counter circuit 254 that generates the MTM
signal, and an inverter 256 that generates MTM. A parallel circuit 258 generatesa delayed one millisecond clock signal MTM+ that is delayed from the MTM clock by about 0.000003289 seconds (i.e., by one half cycle of the 152 KHz CLK
signal).

The choice of a 9.728 MHz clock oscillator 250 and (divide by 65) divider 252 circuits was not arbitrary, but represents the operating frequency region in which 15 quartz crystals can be cut so as to be the most stable over time. Thus the preferred embodiment utilizes inexpensive yet stable components for the clock generator 250.

A digital numerically controlled oscillator (NCO~ 260 generates a linearly 20 increasing signal (modulo 232) called PHASE, which not only increases at a ~requency (19000 + ~/152000) closely matching that of the pilot signal, but alsostarts at a previously computed phase value ~ stored in beacon data table 170.
The PHASE vaiue is converted into a sinusoidal signal PCLK by using the PHASE
value as the address of a SIN() function table stored in a PROM 262. The PCLK
2~ signal is, in essence, a referencQ clock that will be compared by the Phase Detector 144 with the baseband beacon signal.

In parallel with the NCO 260 and PROM 262, an analog to digi~al converter (ADC) 264 samples the baseband pilot signal at a rate o~ 152 KHz. Each sampled pilot 30 signal value is multiplied with a corresponding sample of the PCLK sigr~al bymultiplier 266. During each one millisecond period, 152 consecutive mu~iplication WOg3/0-1378 . PCl/US92/06824 products are received and added to one another by accumulator 268, generating a new phase error value (M) once every millisecond. ;

In both the hardware and software diagrams and equations in this description, the index value M indicates that a new value of the indexed quantity is generated every millisecond.

POWER ESTIMATOR CIRCUIT.
Power estimator circuit 270 also receives the samples of the pilot signal, and 10 generates a power estimation value P(M) once each millisecond. As shown in Figure 5, the power estimator 270 ~irst multiplies each pilot signal sample by itself using a PROM 280 storing an appropriate set of values. The resulting squared values are added to one another for 152 consecutive CLK clock cycles by adder 282 and fli~flop 284, and the most significant eight bits of the resulting total are 1~ loaded into flip~lop 286 to generate the value P(M). The P(M) value is held for one millisecond, providing a period of time in which it can be read and stored by the unit's CPU 150.

The accumulated power value held by the power estimator circuit 270 is cleared 20 once every millisecond by turning off AND gate 288 with the MTM -~ignal, thereby loading a value of zero into the A port of adder 282.

DIGITAL NUMERI~ALLY CONTROLLED OSCILLATOR (NCO.
Referring to Figure 6, the digital NCO 260 works as follows. Only one beacon 25 signai is sampled and analyzed during each 100 miilisPcond period in order toproduce a phase, delta phase, and drift frequency value The digital NCO 260 generates new PHASE values 1~ times per rnillisecond, using a new starting phase value once every millisecond.

WOg3/04378 PCI/IJS92/06824 q~

The theory of operation of the Digital NCO is as follows. The digital NCO 260 generates a PHASE value which is updated, or incremented, about eight times per 19 KHz clock cycle. This PHASE value tracks, as accurately as possible, the phase of a selected beacon signal using a previously measured drift frequency value.

As implemented here, the beacon signal's frequency has two components: (1 ) a 19 KHz component, which causes the phase of the beacon signal to increase .
by ~/4 radians during each clock period of the 152 KHz CLK signali and (2) a delta phase component, which causes the phase of the beacon signal to increase by a~ during each clock period of the CLK signal. In essence, the Digital NCO
circuit 260 computes a value using th~following equation:

PHASE(I) = PHASE(I-1) + 1C/4 + ~ (Eq. 2) where I represents discrete time taken at the 152,000 Hz clock rate.

Prior to the beginning of each one millisecond time slot, the CPU loads a starting phase value ~ into flip-flop 300 and a delta phase value ~ into flip-flop 301.
At the beginning of the one millisecond time slot, marked by MTM=1 and MTM-0, these values are shifted into tli~lops 302 and 304. The delta phase value is added by Adder 306 to 229, which represents a phase shi~t of ~/4 radians. Thus, except during the first clock cycle after MTM=1, the digital 32-bit value at input Port A to Adder 308, is equal to ~/4 + ~.
The value presented to input Port B of Adder 308 is the previously computed PHASE value, stored in flip-flop 312.

At the beginning of each one millisecond time slot, multiplexer 314 selects portA, thereby presenting the phase value in flip-flop 302 to port A of the adder 308.
Flip-flop 312 is cleared by the MTM+ signal one-half clock cycle after MTM

W0 93/04378 ~ 8 PCI/US92/06824 becomes active. As a result, a value of zero is presented to the B port of the adder 308. At the nex~ clock cycle after MTM=1, the output of the adder is equalto ~ + ~, and that value is loaded into fli~flop 312. For each of the ne)n 151 clock cycles of the CLK signal, a value of ~/4 ~ P is added to the previously 5 computed PHASE value and stored in fli~flop 312.

Referring back to Figure 5, the most significant sixteen bits of the computed PHASE value is converted by the PD 144 into a sine waveforrn, and then multiplied with the actual beacon signal to generate a phase error value E. ~:
'~:
PHASE ERROR ACCUMULATOR.
Referring to Figure 7, the main part of accumulator circuit 268 is very simple.
The 32 bit values generated by multiplier 266 (see Figure 4) are accumulated by adder 350 and 11ip-flop 352 for 152 cycles of the CLK signal, with the resulting 15 total value being stored in flip-flop 354 at the end of each one millisecond cycle.
Furthermore, AND gate 356 causes the accumulated value in fli~flop 352 to be cleared at the beginning of each one millisecond cycle and replaced with the next 32-bit value ~rom multiplier 266.

20 The remaining portion of accumulator circuit 268 is a well known overflow detection circuit which operates by comparing sign bits~ In particular, there can -be no overflow if the two inputs to adder 350 have different sign bits. Thus, XOR
gate 360 puts out a ~1~ if the A and B inputs to adder 350 have different signs,which disables AND gate 362.
2~
Furtherrnore, if the previous and current outputs Q~ Addisr 350 have the same sign bit value, the Adder did not overflow. Therefore XOR gate 364 outputs a ,.0u when the previous and current outputs of Adder 350 have the same sign bit value, which disa~les AND gate 362.

WO 93/04378 PCr/US92/06824 2 1 ~ 8 ~ ~:

Overflow is detected when (A) the Sign bits of the multiplier input value and the accumulated phase error value in tlip-flop 352 have the same sign bit value, and (B) the previous and current outputs o~ the Adder 350 have unequal sign bit values. When this happens, AND gate 362 outputs a ~1~, which is stored in ~1i~11Op 366 for the duration of the one millisecond period, and which is then transferred to fli~flop 368 for reading by the CPU 150. The loop filter softwareignores phase error values from each one millisecond period in which the accumulator circuit indicates that an overflow condition was detected.

The phase error value ~ in flip-flop 354 is read by the CPU 150 for use by the loop ~ilter software.

SUMMARY OF PHASE DETECTOR CIRCUIT FUNCTION.
In summary, the PD 144 is tuned to a n~w beacon signal every 100 milliseconds.
1~ After giving the tuner fifty milliseconds to settle, the PD circuit generates three values (estimated signal power, phase error, and a phase error overflow indicator~
once every millisecond for fifty consecutive millisecond periods. These values are passed to the CPU 150.

Using previously measured phase and drift values, the software periodically (i.e., once every 100 milliseconds) updates the predicted phase for the other sPlected beacon signals ~hat are being monitored. As a result, the CPU maintains at all times a phase value ~, delta phase value ~ and drift frequency for all the beacon signals it is using, not just for the one beacon signal that the PD circuit 144 is currently receiving. The procedure for maintaining these phase values will be discussed below, in the section en~itled PHASE TRACKING SOFT\NARE.

w~ 93/04378 2 1 1 4 ~ ~ 8 PC~/US92/06824 PHASE TRACKING ~i~i;~WARE

Referring to Figures ~11, the phase tracking software (used by both the mobile unit and the fixed observer unit) performs the following functions. An automaticgain control routine 400 low pass filters the signal's power (using low pass filter routine 402), and attenuates the phase error signal by a factor inversely proportional to the beacon signal's power.

Another routine 404 uses the phase and drift frequency to remove 38 KHz noise.
10 The resulting filtered error signal ' is summed and averaged over fifty one-millisecond periods by routine 409 to generate error signal E". Loop fil~er 406 then low pass filters this phase error signal " to generate delta phase value and drift frequency value ~'. A phase master accumulator routine 408 computes a new phase value in two torms: PMA, which represents the total 15 amount of phase movement of the ~drift frequency~ portion of the beacon signal since the mobile ùnit was powered on or reset (i.e., excluding phase movement of the 19 KHz ponion of the beacon signal), and ~P, which is the current phase value modulo 2~ radians.

20 After computing these values for the beacon signal which was last processed by the phase detector circuit 144, the CPU ~hen updates ~he phases of the other beacon signals using PHASE update equation Eq. 3:

PMAn(tM) = PMAn(~1) + ~dn (tM 1)-(tM t~ 1) ~Eq. 3) 25 where tM is the time to which the nlh beacon signal phase is being updated t,," is the time corresponding to the phase value stored for ~he n,h beacon signal in the CPU's beacon data table )n'() is the radian drift frequency of the n~h beacon signal WOg3/W378 2~ 8 PCI/US92/06824 -22 - ,, If the stored beacon signal phases are updated on~;e ~ rv 100 milliseconds, then t~ ., = 100 milliseconds.

Operating on the assumption that each of four beacon signals is processed about 5 once every half second, the CPU needs to store certain state variables so as to enable operation of the software filters. These state variables, listed in TABLE
2 herein, are stored in the beacon data table 170. These state variables, excluding those listed in TABLE 1, are the low pass filtered delta phase value ~, the low pass filtered signal power value P', and the low pass filtered drift 10 frequency value ~'.

Eleacon ID ~P P ' ~ ' 096.~; 0.03122 1.00101 0.988 102.1 0.12109 1.23339 0.010 ... ... ... ...
-20 Referring to Figures 8 and 11, the signal power P(M) of the beacon signal is low pass filtered to generate a filtered power value P'(M):

- P'(M) = K-P(M) + (1-K)-P'(M-1) (Eq. 4) 2~ where P(M) is the most recently measured signal power value, P'(M-1) is the previously computed filtered power value, and K is a preselected filter parameter having a value selected for proper bandwidth. A typical value of K is less than 0.1.

30 The amplitude of the sinusoid is recovered from the power measurement by the following equation:

wo g3/04378 2 1 1 4 ~ 4 8 PCI/US~2/06824 -23- .-A'(M) = ~P'(M)/~ -where A'(M) is the amplitude, P'(M) is the sinusoidal power, and ~ is a scaling constant, typically about 56.
The AGC routine 400 attenuates the phase error signal by a factor of olP', where a is an attenuation factor selected for unity loop gain and A' is the low pass tiltered signal amplitude.

Next, routine 404 removes 38 KHz noise from the phase error signal ~. In panicular, when the multiplier circuit 266 in the P0 144 multiplies a computed waveform with the actual beacon signal, two signal components are generated:
a low frequency component with a frequency equal the difference between the two signals, and a high trequency component with a frequency equal to the sum 16 of the two frequencies. After filtering by the accumulator 268, this second component is small, though it is still significant enough to distort the phase error signal 38 KHz compensation routine 404 computes a value G(~',~), corresponding to the unfiltered residual high frequency component of the phase error signal, which is then subtracted by the CPU's software from the phase error 20 signal . G(C~',~) is computed using the following equation:

G( . ~) sin(2WT(N-1)+20 -s~n(2~ ~Eq. ~) 2~ .
where W = 2~-19000 N = 152 T = 1 /152,000 30 The resulting adjusted phase error value ' is averaged by low pass filter routine 409 and then processed by a loop filter routine 406, shown in diagrammatic form wo s3/04378 ~ 6 ~

in Figure 9. This is a ~proportional~ and ~integrated~ signal control routine, in which the delta phase value ~P and drift frequency ~' are periodically updated in accordance with the following equations:

c~'(H) = 1llo2TH~(H) + ~'(H-1) (Eq. 6) ~(H) = 2~CI)o ~(H) + (i)'(H-1) (Eq. 7) where TH jS the hop rate, i.e., the rate at which each station is examined. In 10 the preferred embodiment TH jS nominally 0.~ seconds. ~0 is the natural frequency, in the preferred embodiment~0 is nominally 0.1 radians per second.
is the damping coefficient, which in the preferred embodiment is nominally 0.707.

1~ POSITION DETERMINATION

Periodically, after the CPU has computed the current phase value of the selectedbeacon signals, it deterrnines the mobile unit's current position. In the preferred embodiment, the navigation software 168 is performed concurrently with the 20 mobile unit's signal processing software.

Assuming initial7y that at least four beacon signals were available, position determination is performed as follows.

25 The phasP ~n(t) of each beacon signal at time t, where n is an index iden~ifying particular beacons is represented as:

~)n(t) = ~n(tO) ~ ()n(to)~(t~to~ ~, (t ) (Eq. 8) WO 93/04378 . PC~/US92/06824 where t is time as defined by the fixed observer, to is a specific instance of time chosen by the fixed observer, ~n() iS the phase o~ beacon n at its antenna at to Rn is the distance of the mobile user from the n~h bPacon's antenna ~,() is the wavelength of the pilot tone (approximately 15,789.5 meters) ~n() iS the radian frequency of the nth beacon A problem with Equation 8 is that ~n(t) is not directly observable. In practice,10 the mobile unit can observe phase snly after it has propagated through its receiver, thus being delayed by an amount ~. In other words, the mobile unit can only observe the phase which occurred I seconds in the past. Thus, we modify Equation 8 as follows:

~I)n(t-'t3 = q)n(to) + On(to)-(t~to) ~ ~ )n(tO)~s (Eq. 9j ;~

where T iS an offset in time at the mobile unit due to clock bias or delays within Its receiver.
;
Equation 9 can be re-arranged to yield:

2~ Gn = = ~2to~ Pn(tU) ~ ~n(tO) ~ ~dn(~O)-(tu-tO)}
- ~/(X-Xn)2 + (y yn)2 + (Z z )2 (Eq.10) where tu = t-l is the effective time seen by the user x,y,z is the position o~ the mobile unit ~t time tu xn~yn~zn is the position of the n~h beacon's transmitter WOg3/04378 -26 PCI/US92/068Z4 Repeating Equation 10 for four beacons, and suppressing the equations' dependency upon time for the sake ot clarity, yields a set of four equations with four unknowns (x, y, z and tu). The re.sulting vector of equations can be represented as:
.
G = lG1, G2, G3, G4] (Eq.11) Further, we define a vec~or ot position variables p = [tu, x, y, z].

10 The Jacobian for this set of equations is:
_ aG2 aG2 ~G2 ~G2 J = aG3 ~ aG3 ~G3 ( Eq . 1 ~ ) -~ x ~ z aG4 aG4 aG4 aG4 ~

Next, we iteratively solve the equation:

Pi~, = pj - J '(pj)G(pj), i = 0, 1, 2, .......... (Eq.13) 30 until x, y, z and tu converge.

In our system, with well chosen beacons, the Jacobian matrix is always invertible.
The individual eiements of the Jacobian matrix are numerically computed as foll~ws 3~
aGn _ c = the speed of light - 3 x 10~ m/sec (Eq.14) ~u WO 93/0~378 ~ S ~ 8P~/US92/06824 aGn = ~ (E415) ~x Rn ~Gn = V~Vn (Eq. 16) aGn = ZRZn (Eq. 17) Rn = ~/(x-x~) + (Y-Yn) ~ (Z-Zn) (Eq.18) -where x,y,z is the position of the mobile unit xn~yn~zn is the position o~ the nlh beacon's transmitter 20 To solve ~orthe mobile unit's current position, the mobile unit makes a first guess as to the value of lhe p = ~tu, x, y, z] vector, and assigns that value to pO. Using this value, the sixteen elements of the Jacobian matrix in Equation 12 are computed using Equations 14 through 18. Then the matrix is inverted, and Equation 13 is evaluated to generate a new estimate P1 f the mobile unit's 2~ posi~ion. This process is repeated until the values of t", x, y, and z converge.
As is well known to those skilled in th~ art of numericai methods, ~convergence"is typically de~ined as (t t )2 + (X Xj)2 ~ (Yit.-Yi) + (Zj ,-Zj) < L
or as (tuj"-tu,) ~ , (xj"-x;) < L2X, ~yj~,-y;) ~ L~y~ AND (Zj"-Z;) ~ 1~2 3~ where L1 and L2,-L2z are preselected convergence criteria.

WO 93/04373 PCI'/US!~2/06824 2 ~ 8 As can be seen, the above described computational process not only solves for the mobile unit's position {x,y,z}, it also solves for time tu~ In other words, the mobile unit has sufficient data from its measurements to compute th~ value of time, and tor this reason, it is unnecessary for the mobile unit to explicitly S coordinate its ~clock~ with that of the tixed observer unit. Conceptually, the tu value computed bythe navigation sottware 168 represents the amount of elapsed time between the fixed observer's phase measurements and the mobile unit's phase measurement, plus (A) a time value corresponding to phase shifts caused by propagation of signals through the mobile unit's receiver, and (B) a time value 10 corresponding to the difference between the clock rates of the mobile unit and the fixed observer unit.

It is important to note that the clock rates of the mobile units do not have to be precisely tuned to that of tho fixed observer because any difference in the two 15 clock rates simply becomes a common phase shift for all the received beacons.
NAVIGATION WITH ONLY THREE BEACON SIGNALS.
As shown in Figure 3, the present invention can be used even when only three beacon signals of sufficient quality are available. In that case, the navigation20 software 168 solves only for time, x and y, and does not solve for the mobileunit's altitude. Instead, it is presumed that the mobile unit is at a mathematically fixed altitude. This is based on the assumption that most user's are not as interssted in their altitude as in their latitude and longitude. In the three beacon case, the position vector is p = (tu, x, y}, and the ~lacobian matrix is:

aGl aGl aGl ~tU ax ay J = aG2 aG2 aG2 (Eq .19 ) aG3 aG3 aG3 3~ at~, ax ay wo 93/04378 ~ 8 PCI`/US92/06824 Fnuation 13 remains unchanged, except that the vectors and matrix now have fewer elements, and the iterative process for computing the mobile unit's position remains the same as described above.

5 USER INTERFACE.
The present invention does not require any specific user interface for communicating the computed mobile unit's position. Typically, in motor vehicle applications, the mobile unit's current position will be displayed by indicating a position on a map that is displayed on a computer display device. However, 10 other types of user interfaces may be used, depending on the user's requirements.
.

INITIALIZING THE NAVIGATION ROUTINE.
One problem with the invention as described so far is that the system needs 15 an initial estimate o~the mobile unit's position that is correct to within a ~ew miles.
Given such an estimate, the system as described will use the estimate as an initial guess, and then will quickly converge to the correct user position.

However, the solution to the navigation equations presented above is not 20 guaranteed to be unique. If the initial estimate of position is too far in error, the system may converge to a solution that does not represent true user position but is othervvise consistent with observations. The fundamental cause of ~his behavior is the fact that the wavefonn being tracked (i.e., a 19000 Hz sinusoid)has a wavelength of only 15.6 kilometers (about ten miles). That is, once every 2~ 1~.6 Km, the pilot tone repeats. Without extra inform2tion, the system is incapable of determining which cycle of the pilot tone is being received by the mobile unit. Thus, there is a potential ambiguity problem.

There are several method for obtaining the information required to resolve this 30 ambigulty.

WO 93/04378 ~ 6 ~ 8 PCr/US92/06824 In som~ :mplementations, such as systems in which a human operator is always present (e.g., in an automobile navigation system), the system can be programmed simplyto askthe userto specifythe system's approximate position each time that the system is powered on. Since the wavelength of the bea~ons is about ten miles, the user needs only to specify the system's position within about five miles. This may be done by displaying a map on a touch screen display and asking the user to indicate his/her position by touching the corresponding position on the screen. For a system used over a wide geographic region, this process could take several steps, tor example starting with a national 10 map in which the user specifies the general region of the system, moving thento a map covering a region the size of one or two states, and then moving to a more local map for a final pinpointing of the user's position within five miles.

Given this approximate starting position, the navigation system will collect phase 15 measurement data and then execute its navigation sottware 168 to precisely determine the sys~em's position with a high degree of accuracy. In one preferredembodiment, the mobile unit's last computed position is stored in non-volatile memory (such as an EEPROM), and that position is used as the pO value the first time the mobile unit computes its position after power up or reset. Position 20 need be re-entered by the user only when the mobile unit's position changes significantly while the mobile unit is off.

In the preferred embodiment, the mobile unit's position is typically recomputed about every 0.~ seconds. Lower cost implementations, using slower 2~ microprocessors, might recompute position less often. Even if the mobile unitis moving at a rather high rate of speed, such as 1000 Kilometersthour in an airplane, the previously computed position will still be relatively close to the new position, and therefore the computation will typically converge after a few iterations of the above describe computation.

WO 93/04378 ~ 6 ~ 8 PCI'tUS92/06824 Another method of rti^olving the ambiguity problem is to take advantage of otheNvise minor imperfections in the beacons to set up mathematical ~virtual beacons~ with ambiguity ranges ~ar larger than 10 miles. This can be accomplished because, though the frequency of the beacons is nominally 19000 5 Hz, FCC regulations allow a i 2 Hz deviation ~rom this ideal. Observed deviations are typically within ~ 1.2 Hz from the ideal. -Consider a set o~ beacons taken pair-wise. Let these bsacons have slightly different frequencies. Considering only two such beacons for the moment, a 10 comparison of the phases o~thesetwo beacons results in a beat ~requency whichis a ~unction of the trequency separate o~ the two. For e^xample, if the ~requency o~ the two stations is separated by 2 ~z, then the difference between the two phases will repeat at most once every half second. This sets up an effective virtual beacon with a wavelength equal to the speed o~ light divided by the 1~ separation in ~requencies. Again using a 2 Hz separation as our example, the effective wavelength is approximately 150,000 Kilometers. Making the rather safe assumption that we are within a Sew hundred kilometers of the beacons, that is, that we are on the surface of the planet and within range of both beacons, we are clearly w~hin the first region of repetition ~or this virtual beacon, and reject 20 subsequent repetitions as possibilities Plainly, four beacons can be used to set up four different virtual beacons in this fashion, and these virtual beacons can be used to solYe lor user position in almost the same fashion as is used ~or the actual beacons as described above 2~ ~he longer wavelength implies a decrease in solution accuracy, but our purpose is only to obtain an approximate solution as an initial estimate ~or use with the navigation equations that use actual beacons Yet another technique to obtain the required extra information needed to resolve30 ambiguity involves reducing the degrees of ~reedom in the system while wO g3/04378 ~ 6 Q~ ~ PCl`/US92/06824 , - 32 ~
maintaining the number of be~n.s used, or alternatively adding beacons.
Degrees of freedom are reduced by using extemal information, notably map data.
Wffl map data, for example, one can independently obtain altitude information as a function of x and y position. Using this infonnation, solutions which do not 5 exhibit the proper x, y and z relationships can be discarded. Because of the large wavelength (ten mile) of the preferred waveform, two or more solutions which exhibit the same x, y and z relations are highly unlikely.

A similar approach, herein called the quantized map approach, divides the local 10 region into one or more sections and describes the z axis by using a single number representing the average local altitude for that section. With sufficiently fine quantization, consistent x, y and z solutions are rendered unlikely except for the unique position at which the mobile unit is tound.

15 Yet another technique uses extra information regarding changes in the observed phases as a function ot changing system geometry. System geometry is changed either by movement of the mobile unit or by using different beacons, which is tantamount to ~beacon movement~. User movement allows the system to examine the changes in observed phases as a function of changing position.
20 There may be many soiutions consistent with a set of phase observations, but only one will remain consistent in the presence o~ user movement. After observing user movement for a sufficient period of time, only one solution, the mobile unit's true position, will survive. This is particularly suitable ~or systems that are mountPd on moveable vehicles but which either lack a user interface 25 or in which a user may not be able to specify the system's starting pos~ion.

An equivalent approach is to add yet another beacon to take a different constellation of beacons while keeping the number of degrees of freedom the same. If, for example, we wish to solve for x, y, z and time, then a minimum 30 of four beacons are required to solve for all quantities. An infinite number of WO g3/04378 . PCI`/US92/06824 solutions are available, but this numbe~ .- rendered finite using the safe assumptions that we are near the surface of the planet and within range oS the beacons that the mobile unit receives. This ~inite number o~ solutions can be reduced to a single unique answer by using a fifth beacon to resolve which 5 solution is consistent. Plainly, this technique can be extended to use even more beacons. Only one solution will remain consistent, the true mobile unit position.

FIXED POSITION OBSERVER UNIT

Referring to Figures 1, 2, 3 and 12, the fixed position observer unit 110 has the same hardware as shown in Figure 2 for a mobile receiver unit 120, except that the fixed position unit 1 10 is coupled t~a transmitter for the purpose of sending reference data to the mobile units, as shown in Figure 1 . Furthermore, the signal processing software (i.e., the loop filter, phase master accumulator, low pass 1~; fiiter, and phase update routines) for the fixed position unit is basically the same as the mobile unit's software.

The navigation routine 168 of the rnobile unit is replaced by a routine for converting the detected phase of each beacon signal into a phase value at the 20 position of the beacon's transmitter antenna at a selected reference time to. This is simply a matter of computing a phase shift based on the distance between the fixed observer unit 110 and the transmmer antenna, and the difference between the reference time to and the time that the phase was measured. The required computation is shown by Equation 20:
2~
~n(t) = ~n(t0) + ~ (Eq. 20) where ~)n() is the phase observed by the fixed observer q)n() is the computed phase of the nth beacon signal at its antennae WO g3/04378 PCI~/US92/06824 4 8 ~ ~

Note that ~ is the wavelength of the received signal, where ,. - ~'(19,000 + c~'/2~) and ~' is the dritt ~requency measured by the fixed observer ~see Figure 8).

Referring to Figure 12, the main control routine for the fixed observer operatesas follows. At steps 452 and 454, the fixed observer unit scans all the available FM stations to find those with sufficient signal quality for further processing.
For each station with acceptable signal quality (strength and antenna position),the fixed observer measures the phase of its beacon signal (i.e., its 19 KHz pilot 10 tone) at the fixed observer's position (step 4~6). Then it computes the phaseof that beacon signal at the transmittin~g antenna's position, at a selected time to~ using Equation 20, above, and stores the adjusted phase value in a beacon data table (steps 458 and 460). The data values currently in the beacon data table are continuously broadcast to the mobile units. It has been observed that 1~ the frequencies and phases ot the pilot tones of FM stations are very stable,and that updating the reference data ~or each beacon signal once eve~ ~ew seconds is more than sufficient to maintain an accurate positioning system.

Each mobile unit uses the reference data from ~he fixed observer to locally 20 regenerate a set o~ beacon signals with coordinated phases, closely replicating the actual beacon signals at positions o~ their antennae. Thus the role of the fixed observer is to provide data that enables the mobile units to locally regenerate accurate replicas of the beacon signals. Since the relative phases of the beacon signals are known, the beacon signals are effectively coordinated 2~ beacons. The mobile unit mixes each of these regenerated beacon signals with the corresponding received beacon signai, thereby producing a high accuracy phase value ~or each beacon signal. The navigation routine then computes the mobile unit's absolute position, in the coordinate system of the beacon antennae, based on these phase values.

wo 93/W378 ~ 1 1 4 6 ~ 8 PCI/US92/06824 - 35 ~
ALTERNATE EMBODIMENTS

The beacon signals used by the present invention need not be sinusoidal waveforms. For instance, a square wave or pulsed beacon signal could be used, requiring only that the system's phase detector circuit measure the phase of transitions in the received beacon signal. Other beacon signal waveforms could be used, so long as the phase of the received beacon signal can be resolved accurately by both the fixed position observer and the mobile units.

10 In an alternate embodiment, the fixed observer unit is programmed to transmitnot only phase values for each beacon signal, but also a drift frequency value.
The drift frequency value is then use~ by the mobile unit in its computations, in place of the drift frequency measured by the mobile unit, so as to eliminate any Doppler shift in the frequency value caused by movement of the mobile unit.

It isnotedthatwhilethe preferredembodiment usestransmittersthat are atfixed positions, the present invention could be used with mobile or orbiting transmitters, so long as the positions of the transmitters can be precisely determined at any specified point in time.
The present invention cou!d be used wi~h any stable radio station sub-carrier signal, preferably having a frequency between 1 KHz and 1Q0 KHz.

While the present invention has been described with reference to a few specific 2~ embodimen~s, the description is illustrative of the inven~ion and is not to be construed as limiting the invention. Various modifications may occur to those skilled in the art without departing from the true spirit and scope of the invention as defined by the appended claims.

Claims (22)

WHAT IS CLAIMED IS:

1. A positioning system, comprising:
a multiplicity of transmitters, at known fixed locations, each of which transmits a beacon signal having a phase that is un-synchronized with the phasesof the beacon signals of the other transmitters; each of said beacon signals having a frequency which is approximately equal to a predetermined target frequency;
a first receiver, at a known location relative to the locations of said multiplicity of transmitters, which (A) receives said beacon signals, (B) determines the phases and frequencies of said beacon signals at times determined with reference to a first time reference signal, and (C) broadcasts phase and frequency data representing said beacon signal phases and frequencies; said broadcast phase and frequency data including time data representing when, relative to saidfirst time reference signal, said beacon signals attained said beacon signal phases;
a second receiver, at an unknown location, which receives said broadcast phase and frequency data from said first receiver and at least three of said beacon signals, said second receiver including a local clock generator, un-synchronized with said first time reference signal, that generates a second time reference signal, phase detection circuitry for detecting, at times determined with reference to said second time reference signal, the phases and frequencies of said beacon signals at the location of said second receiver, and computation means for computing, based on said broadcast phase and frequency data and said detected phases and frequencies, coordinate values defining said second receiver's position and a time adjustment value for bringing said second time reference signal into alignment with said first time reference signal, said time adjustment value further relating said detected phases to saidbroadcast phase data in accordance with a predefined set of simultaneous equations.

2. The positioning system of claim 1, said second receiver including a single, frequency scanning radio signal receiver for sequentially receiving selected beacon signals at a multiplicity of distinct carrier frequencies;
said phase-detection circuitry including a single digital phase-locked loop, coupled to said frequency scanning radio signal receiver, which receives said beacon signal and generates a phase error signal;
said computation means including phase computation means, coupled to said phase-detection circuitry, for integrating said phase error signal while said second receiver is receiving each said selected beacon signal, and generating a distinct phase and frequency value for each said selected beacon signal;
said computation means further including software means for performing a predefined position computation using said generated phase and frequency values and said broadcast phase and frequency data received from said first receiver.

3. The positioning system of claim 1, said second receiver including memory means storing transmitter position data indicating the positions of said multiplicity of transmitters;
said computation means including software means for performing a predefined position computation using said stored transmitter position data to generate a position value for said second receiver which is consistent with saidstored transmitter position data.

4. The positioning system of claim 1, wherein said beacon signal transmitted by each of said transmitters comprises an FM radio station sub-carrier signal having a frequency of approximately 19 KHz.

5. The positioning system of claim 1, wherein said beacon signal transmitted by each of said transmitters comprises a sub-carrier signal on a radio station carrier signal, said sub-carrier having a frequency between 1 KHz and 100 KHz.

6. The positioning system of claim 1, said computation means including (A) phase adjustment means for adjusting said detected phases in accordance with said detected frequencies so as to represent said detected phases at a common time relative to said second time reference signal prior to computing said coordinate values and time adjustment value; and (B) memory means for storing transmitter position data indicating positions of said transmitters with respect to a predefined coordinate system; and (C) means for performing a predefined position computation using said stored transmitter position data to generate a position value in said predefinedcoordinate system for said second receiver.

7. The positioning system of claim 1, said first receiver including a first clock generator that generates said first time reference signal.

8. In a positioning system which makes use of a multiplicity of transmitters, at known fixed locations, each of which transmits a beacon signal having a phasethat is un-synchronized with the phases of the beacon signals of the other transmitters; each of said beacon signals having a frequency which is approximately equal to a predetermined target frequency;
the combination comprising:
a first receiver, at a known location relative to the locations of said multiplicity of transmitters, which (A) receives said beacon signals, (B) determines the phases and frequencies of said beacon signals relative to each other, and (C) generates a first set of phase and frequency data representing said beacon signal phases; said first receiver including a first clock generator that generates a first time reference signal; said broadcast phase and frequency data includingtime data representing when, relative to said first time reference signal, said beacon signals attained said beacon signal phases;
a second receiver, at an unknown location, which receives at least three of said beacon signals, said second receiver including a local clock generator, un-synchronized with said first clock generator, that generates a second time reference signal, and phase detection circuitry for detecting, at times determined with reference to said second time reference signal, the phases and frequencies of said beacon signals at the location of said second receiver; and computation means, coupled to said second receiver for receiving from said second receiver values corresponding to said detected beacon signal phases and frequencies, coupled to said first receiver for receiving said first set of phase and frequency data, and including means for computing, based on said first set of phase and frequency values and said detected beacon signal phases and frequencies, coordinate values defining said second receiver's position and a time adjustment value for bringing said second time reference signal into alignment with said first time reference signal, said time adjustment value further relating said detected phases to said broadcast phase data in accordance with a predefined set of simultaneous equations.

9. The combination of claim 8, said second receiver including a single, frequency scanning radio signal receiver for sequentially receiving selected beacon signals at a multiplicity of distinct carrier frequencies;
said phase-detection circuitry including a single digital phase-locked loop, coupled to said frequency scanning radio signal receiver, which receives said beacon signal and generates a phase error signal;
said computation means including phase computation means, coupled to said phase-detection circuitry, for integrating said phase error signal while said second receiver is receiving each said selected beacon signal, and generating a second set of phase and frequency values comprising a distinct phase and frequency value for each said selected beacon signal;
said computation means further including software means for performing a predefined position computation using said second set of phase and frequency values and said first set of phase and frequency data received from said first receiver.

10. The combination of claim 8, said computation means including memory means storing transmitter position data indicating the positions of said multiplicity of transmitters; and position computation means for performing a predefined position computation using said stored transmitter position data to generate a position value for said second receiver which is consistent with said stored transmitter position data.

11. The combination of claim 8, said second receiver receiving at least four of said beacon signals and detecting the phases and frequencies of said at least four beacon signals at thelocation of said second receiver at times determined with reference to said second time reference signal; and said coordinate values computed by computation means comprising a three-dimensional set of coordinate values defining said second receiver's position with respect to a predefined three-dimensional coordinate system.

12. A method of determining a mobile unit's position with respect to the positions of a multiplicity of transmitters that are positioned known fixed locations, wherein each transmitter transmits a beacon signal having a phase that is un-synchronized with the phases of the beacon signals of the other transmitters;each of said beacon signals having a frequency which is approximately equal to a predetermined target frequency;
the steps of the method comprising:
at a known location relative to the locations of said multiplicity of transmitters, receiving said beacon signals, determining the phases and frequencies of said beacon signals at times determined with reference to a firsttime reference signal, and broadcasting data representing said beacon signal phases and frequencies, said broadcast data including time data representing when, relative to said first time reference signal, said beacon signals attainedsaid beacon signal phases;

at a mobile unit's location, receiving said broadcast phase data and at least three of said beacon signals, generating a second time reference signal that is un-synchronized with said first time reference signal, detecting, at times determined with reference to said second time reference signal, the phases and frequencies of said beacon signals at the location of said mobile unit, and computing based on said broadcast data and said detected phases and frequencies, coordinate values defining said mobile unit's location and a time adjustment value for bringing said second time reference signal into alignment with said first time reference signal, said time adjustment value further relating said detected phases to said broadcast phase data in accordance with a predefined set of simultaneous equations.

13. The method of claim 12, including storing transmitter position data indicating the positions of said multiplicity of transmitters;
said computing step including performing a predefined position computation using said stored transmitter position data to generate a position value for said mobile unit which is consistent with said stored transmitter position data.

14. The method of claim 12, at said mobile unit's location, receiving at least four of said beacon signals;
said detecting step including detecting the phases and frequencies of said at least four beacon signals at the location of said mobile unit at times determined with reference to said second time reference signal; and said computing step including computing a three-dimensional set of coordinate values defining said mobile unit's location with respect to a predefined three-dimensional coordinate system.

15. The method of claim 12, including generating said first time reference signal at said known location.

16. A method of determining a mobile unit's position with respect to the positions of a multiplicity of transmitters that are positioned known fixed locations, wherein each transmitter transmits a beacon signal having a phase that is un-synchronized with the phases of the beacon signals of the other transmitters;each of said beacon signals having a frequency which is approximately equal to a predetermined target frequency;
the steps of the method comprising:
at a known location relative to the locations of said multiplicity of transmitters, receiving said beacon signals, determining the phases and frequencies of said beacon signals at times determined with reference to a firsttime reference signal, and generating phase and frequency data representing said beacon signal phases and frequencies, said generated phase and frequency data including time data representing when, relative to said first time reference signal, said beacon signals attained said beacon signal phases;
at a mobile unit's location, receiving said at least three of said beacon signals, generating a second time reference signal that is un-synchronized with said first time reference signal, and detecting, at times determined with reference to said second time reference signal, the phases and frequencies of said beacon signals at the location of said mobile unit; and receiving said generated phase and frequency data, receiving values corresponding to said beacon signal phases and frequencies detected at said mobile unit's location, and computing based on said received phase and frequency data and said phases and frequencies detected at said mobile unit's location, coordinate values defining said mobile unit's location and a time adjustment value for bringing said second time reference signal into alignment with said first time reference signal, said time adjustment value further relating said phases detected at said mobile unit's location to said received phase data in accordance with a predefined set of simultaneous equations.

17. The method of claim 16, including storing transmitter position data indicating the positions of said multiplicity of transmitters;
said computing step including performing a predefined position computation using said stored transmitter position data to generate a position value for said mobile unit which is consistent with said stored transmitter position data.

18. The method of claim 16, at said mobile unit's location, receiving at least four of said beacon signals;
said detecting step including detecting the phases and frequencies of said at least four beacon signals at the location of said mobile unit at times determined with reference to said second time reference signal; and said computing step including computing a three-dimensional set of coordinate values defining said mobile unit's location with respect to a predefined three-dimensional coordinate system.

19. The method of claim 16, including generating said first time reference signal at said known location.

20. A positioning system, comprising:
a multiplicity of transmitters, at known fixed locations, each of which transmits a beacon signal having a phase that is un-synchronized with the phasesof the beacon signals of the other transmitters; each of said beacon signals having a frequency which is approximately equal to a predetermined target frequency;
a first receiver, at a known location relative to the locations of said multiplicity of transmitters, which (A) receives said beacon signals, (B) determines the phases and frequencies of said beacon signals at times determined with reference to a first time reference signal, and (C) broadcasts phase and frequency data representing said beacon signal phases and frequencies; said broadcast phase and frequency data including time data representing when, relative to saidfirst time reference signal, said beacon signals attained said beacon signal phases;

? second receiver, at an unknown location, which receives at least three of said beacon signals, said second receiver including:
a local clock generator, un-synchronized with said first time reference signal, that generates a second time reference signal; and phase detection circuitry for detecting, at times determined with reference to said second time reference signal, the phases and frequencies of said received beacon signals at the location of said second receiver; and position computation means which receives said broadcast phase and frequency data from said first receiver and said phases and frequencies detectedby said second receiver, said position computing means including means for computing, based on said broadcast phase and frequency data and said detected phases and frequencies, coordinate values defining said second receiver's position and a time adjustment value for bringing said second time reference signal into alignment with said first time reference signal, said time adjustment value further relating said detected phases to said broadcast phase data in accordance with a predefined set of simultaneous equations.

21. The positioning system of claim 20, said first receiver including a first clock generator that generates said first time reference signal.

22. The positioning system of claim 20, said second receiver receiving at least four of said beacon signals and detecting the phases and frequencies of said at least four beacon signals at thelocation of said second receiver at times determined with reference to said second time reference signal; and said coordinate values computed by computation means comprising a three-dimensional set of coordinate values defining said second receiver's position with respect to a predefined three-dimensional coordinate system.