US 3883900 A
A prosthetic arm has an upper arm member having a stump receiving socket, a forearm member and an elbow unit secured to the upper arm member and pivotally connected to the forearm member. A drive, housed within the elbow unit, includes a reversible direct current permanent magnet torque motor and a transmission including a planetary gear reduction unit, a reverse locking clutch, and a plano-centric unit and the transmission is connected to the forearm member with its output shaft part of the pivotal connection therewith. The forearm member houses a battery pack and the circuitry by which the motor is operated in either direction in response to electromyographic signals that may be picked up from the biceps and triceps by electrodes when attached to the stump and processed to drive the motor in a direction and at a rate dependent on the dominant EMG signals. The locking clutch is operable to hold the arm flexed against a predetermined load and the elbow unit also houses a tachometer to provide a feedback to modify the power supplied to the motor to enhance the controlability of the amputee of flexing velocities.
Claims available in
Description (OCR text may contain errors)
United States Patent J erard et al.
[ BIOELECTRICALLY CONTROLLED PROSTHETIC MEMBER Inventors: Robert B. Jerard, Brattleboro, Vt.;
Cord W. Ohlenbusch, Hopkinton, Mass Liberty Mutual Insurance Company, Boston, Mass.
Filed: Sept. 7, 1973 Appl. No.: 395,236
US. Cl 3/l.1; 3/l2.3 Int. Cl A6lf l/06; A6lf l/OO Field of Search 3/1.1, 1212.3
References Cited UNITED STATES PATENTS 1/1971 Ohlenbusch et al. 3/1.!
Primary Examiner-Ronald L. Frinks  ABSTRACT A prosthetic arm has an upper arm member having a stump receiving socket, a forearm member and an elbow unit secured to the upper arm member and pivotally connected to the forearm member. A drive, housed within the elbow unit, includes a reversible direct current permanent magnet torque motor and a transmission including a planetary gear reduction unit, a reverse locking clutch, and a plano-centric unit and the transmission is connected to the forearm member with its output shaft part of the pivotal connection therewith. The forearm member houses a battery pack and the circuitry by which the motor is operated in ei' ther direction in response to electromyographic signals that may be picked up from the biceps and triceps by electrodes when attached to the stump and processed to drive the motor in a direction and at a rate dependent on the dominant EMG signals. The locking clutch is operable to hold the arm flexed against a predetermined load and the elbow unit also houses a tachometer to provide a feedback to modify the power supplied to the motor to enhance the controlability of the amputee of flexing velocities.
21 Claims, 14 Drawing Figures iZUENTEU N- SHEET 10F 8 PMEMEDH 3.883.900
SHEET 2 OF 8 FIG. 4
1 BIOELECTRICALLY CONTROLLED PROSTHETIC MEMBER BACKGROUND OF THE INVENTION For several years, it has been apparent that bioelectric control of prostheses would be highly advantageous as such control could be similar to the control of the corresponding body section lost through amputation. The muscles of the stump produce electrical signals which can be sensed directly from the surface of the skin. Such signals are called electromyographic signals and are herein often referred to as EMG signals.
In U.S. Pat. No. 3,557,387 dated Jan. 26, 1971, a joint prosthesis is detailed in which such signals are effectively utilized to control the flexing of a forearm member with the advantage generally referred to above. The forearm member housed the motor, the transmission, an electromechanical brake, and the circuitry while the current demands were such that the power source had to be external due to its size and weight.
In general, prosthetic arms in which flexing is effected by EMG signals have had transmissions that were relatively heavy and bulky and additionally were not as smooth and quiet in operation as desired. Evaluations and suggestions by amputee users have pointed to a need for a decrease in weight and noise and a smaller, less cumbersome battery pack.
It should be here noted that proposals have been made to use plano-centric drives in prosthetic arms. See Livingston S.M., D. I. Crecraft, Design of an Artificial Elbow; an Electromechanical Solution, Control of Artificial Limbs, the Institution of Mechanical Engineers, London 1968.
While plane-centric drives offer advantages in size and strength, they have been excessively noisy at high speeds and run very roughly at low speeds.
THE PRESENT INVENTION The general objective of the invention is to provide a jointed prosthesis of appropriate weight and weight distribution with flexing responsive to EMG signals derived from the stump to which the prosthesis is attached and with the actuating mechanism, the circuitry, and the power source all contained within the prosthesis.
This general objective is attained by providing a prothesis with a joint unit housing a motor and transmission. The joint unit is secured to one member with the output shaft of the transmission connected to the other member in a manner such that said other member is flexed in a direction dependent on the direction in which the motor is operating with the transmission providing a substantial gear reduction between the motor and the driven shaft and including a reverse locking clutch to enable the prosthesis to be locked in a flexed Another objective of the invention is to provide a locking clutch that overcomes problems of previous brakes that may be summarized as over-riding load chatter, an objective attained by providing a clutch that becomes operative as a brake by the jamming of rollers as a consequence of a load but with instant locking prevented by including in the clutch frictionally engaged surfaces. The tachometer controlled velocity feedback is operative to place the motor in operation at a rate commensurate with that wanted by the ampu- Other objectives of the invention are to enable the amputee to have better control of flexing velocities and to enable starting friction to be readily overcome, both objectives attained by the use of a tachometer coupled to the motor and providing a velocity feedback effecting appropriate modification of the power input to the motor.
Another objective of the invention is to provide for further conservation of power by the use of limit switches arranged to prevent overtravel of the flexed member of the prosthesis in either direction and to permit the prosthesis to remain in its fully extended or its fully flexed position without the motor drawing battery current.
BRIEF DESCRIPTION OF THE DRAWINGS In the accompanying drawings, a preferred embodiment of the invention is shown, and
FIG. 1 is a side view of a prosthetic arm in accordance therewith;
FIG. 2 is an exploded view of the elbow unit showing the components housed therein;
FIG. 3 is an exploded view of the transmission showing its components;
FIG. 4 is a section, on an increase in scale, taken lengthwise of the elbow unit along the axis of its connection with the forearm section;
FIG. 5 is a section taken approximately along the indicated line 55 of FIG. 4 showing the planetary gear- FIG. 6 is a section taken approximately along the indicated line 6-6 of FIG. 4 showing the reverse locking clutch;
FIG. 7 is a section taken approximately along the indicated line 77 of FIG. 4;
FIG. 8 is a section taken approximately along the indicated line 88 of FIG. 4;
FIG. 9 is a block diagram of the mechanical and electrical components;
FIG. 10 is a schematic view of the EMG amplifiers and the EMG signal processing;
FIG. 11 is a like view of the signal summer and the velocity feed back amplifier;
FIG. 12 is a like view of the pulse width modulator FIG. 13 is a schematic view of the battery filter; and
FIG. 14 is a like view of the power amplifier.
An artificial arm in accordance with the invention includes, see FIG. 1, a conventional upper arm section generally indicated at 15 with parts of its attaching harness indicated at 16 and 17. The arm section 15 has an end plate 18 at its distal end.
A forearm section, generally indicated at 19, has a shell 20, shown only in phantom, enclosing a frame which consists of sides 21 and 22 in support of a plate 23 for circuit boards indicated generally at 24 and 25 and a front end wall 26 to which a holder 27 is attached, the holder 27 providing support for a terminal device, such as the conventional hook 28. The holder 27 is detachably attached to the end wall 26 and its length is determined by the forearm length appropriate for each amputee. At their other ends, the sides 21 and 22 have transverselyaligned holes 29 and 30, respectively, the hole 29 round and the hole 30 square. Adjacent said other ends, the frame sides 21 and 22 support a holder 31 of U-shaped cross section for the support of the battery pack 32'.
The arm sections 15 and 19 are interconnected by a prosthetic elbow unit generally indicated at 33 in FIG. 1 and, as shown inFIGS. 2 and 4 it has a housing formed in two parts, the part 34 and the part 35 clamped together by screws 36. A reversible direct current, permanent magnet torque motor, a tachometer,
generally indicated at 37 and 38 respectively, and a transmission later to be detailed are within the housing. The housing parts 34 and 35 have surfaces 34A and 35A that are coplanar when the housing is assembled and seat against the upper arm plate 18 with the housing surface 34A provided with a threaded stem 34B by which the elbow unit 33 is attached to the plate 18 of the upper arm section 15. The housing part 35 has an internal, annular recess 35B for an internally toothed rim gear or circular spline 39 clamped between the housing parts 34 and 35 and held against turning relative thereto as by the screws 36.
The drive shaft 40 of the transmission is rotatable in the end wall of the housing part 34 and has an outwardly disposed square hub 40A fitting the square hole 30 of the frame side 21 which is locked thereto by a screw 41 and an interposed washer 42. The housing part 35 has anchored therein an outwardly disposed axial, internally threaded insert 43 extending through the hole 29 of the frame side 22 and locked thereto by a screw 44. On both sides of the unit 33 there is a stop 45 engageable by the proximate frame end to limit the extent to which the forearm section 19 can be flexed or straightened. In practice, the arc through which the forearm section may swing is 130.
The drive shaft 40 has a flange 40B at its inner end clamped to the end wall of the casing 46 of the motor 37 and the interposed end wall of the flexible spline 47 by screws 48. At its other end, the spline 47 has a series of lengthwise teeth or splines 49 meshing with the rim gear or fixed spline 39 in a manner subsequently detailed.
Within the casing 46 is the stator 50, the brush ring 51 and the rotor 52 of the motor 37 with the ends of the rotor shaft 53 supported by ball bearing units 54, one fixed in the end wall of the motor casing 46 and the other in its end cap 55 which with an internally toothed ring gear 56 and an outer clutch housing 57 are clamped to the open end of the casing 46 by screws 58.
The end of the drive shaft 40 and the end wall of the motor casing 46 have sockets defining a chamber 59 for the tachometer 38 whose coupler 60 is entrant of a bore 53A on the proximate end of the motor shaft 53 and establishes a rigid coupling therewith.
The planetary gear reduction unit or section of the transmission, see FIGS. 4 and 5, includes the ring gear 56, a clutch housing 61 having a flange 61A clamped between the ring gear 56 and the outer clutch housing 57. A rotatable circular input plate 62 has a series of gears 63 rotatably mounted thereon each in mesh with the drive gear 64 fixed on the end of the drive shaft 53 of the motor 37 that extends through the end cap 55 and with the fixed ring gear 56 so that the input plate 62 rotates at a desirably reduced rate in either direction in which the motor is operated. In practice and by way of example, the ratio is 4.33:1. The input plate 62 has a stub shaft 65 entrant of a socket in the shaft 66 of a square output cam 67 which is a free fit within the housing 61, the shaft 66 running in ball bearing units 68 in the hub of the clutch housing 61.
The input plate 62 has two diametrically opposed pairs of drive portions 69 and 70, see FIG. 6, that are spaced apart with their outer surfaces arcuate and litting freely within the housing 61 and their inner surfaces parallel to and receiving between them opposite sides of the output cam 67 and spaced relative thereto to provide a driving connection therewith. The arcuate extent of the portions 70 is greater than that of the portions 69 and a roller 71 is freely confined between the drive portions 69 and 70 of each pair.
The transmission also includes a low ratio planocentric drive, see FIGS. 7 and 8, shown as a Harmonic Drive generally of the type disclosed in US. Pat. No. 2,906,143 and manufactured by USM Corp. of Boston, Mass. The drive has a flanged output hub 72 pinned to the shaft 66 and transversely slotted as at 73 to loosely receive the internally disposed, transversely aligned splines 74 in the bore of a coupling 75 having a flange 76 and held in place by a retaining ring keeper 77 on the end of the hub 72. The coupling 75 has oppositely disposed external splines 78 each to fit the recess between the ends of circumferentially extending internal ribs 79 at one end of the wave generator plug 80 whose outer surface is elliptical as is apparent from FIG. 7. A ball bearing unit 81 has its races so made that the open end of the flexible spline 47 is thus made sufficiently elliptical to cause its splines 48 to move into engagement with the internally toothed rim gear or circular fixed spline 39 in two opposed zones as is characteristic of such a drive, which in practice, provides a reduction in the order of 64:1.
With this transmission, when the motor 37 is operated to raise the forearm section 19, the input cam portion 69 drives to the output cam 67 counterclockwise with the rollers 71 held so that they cannot become jammed against the housing 61. Similarly, when the motor 37 is reversed to lower the forearm section 19, the input cam portion 70 drives the output cam 67 clockwise, after the input cam portion 69 first dislodges the rollers 71 if they have become jammed. If the arm is driving the motor 37 counterclockwise, the output cam 67 drives the input cam portion 70 and the rollers 71 cannot jam.
If, however, the forearm 19 is in a partly flexed position, whether or not in support of a load, the output cam 67 is biased in a clockwise direction thereby forcing the rollers 71 to become jammed against the housing 61. As stated earlier, the housing flange 61 is clamped against the ring gear 56 by the outer clutch housing 57 thereby to provide frictional resistance against the turning of the housing 61 unless the load exceeds a predetermined value and should it move, the tachometer 38 senses such movement to bring the motor 37 back in service. i
'The main problem with reverse-locking clutches occurs with an over-riding load. For example, if the amputee wishes to lower a heavy load, two problems are present. The first of these is that when the input cams 69 unjam the rollers 71, the forearm section 19 instantly starts to take off" in the down direction with both the load and the motor 37 driving it. This effect is minimized by virtue of the fact the jamming angle is not less than l2.l4 nor more than 15.l4 which keeps the motor torque necessary to unjam the rollers 71 to a minimum. The tachometer 38 then provides the velocity feedback necessary to reduce the power to the motor 37 to prevent the take off.
The second problem results from the fact that the output cam 67, being unsupported and free to move in the-same direction as the input cams are travelling, now jams the rollers 71 back against the housing 61. The input cams 69 immediately unjam the rollers 71 and the repeated jamming and unjamming results in a chatter effect as the amputee lowers the load. This effect is made minimal by the fact that the drive flange 61A is frictionally held by the pressure of the flange of the outer housing 57 as a disc brake, permitting a degree of slippage with chatter eliminated in the case of light to moderate loads and minimized in the case of heavy loads.
The overall functioning of the electronic circuitry is essentially the same as described in U.S. Pat. No. 3,557,387 but is quite different in order to achieve several additional desirable features. These features are lower power consumption, better common mode rejection, the incorporation of several deadbands whose use is subsequently described and better reliability. The overall functioning of this embodiment is described only generally in connection with, FIG. 9. The batteries of the pack 32 are of 6 volts and of the nickel cadmium type and provided with a columbmeter 82 to allow the battery or batteries to be quickly recharged and the battery leads to the power amplifier 83 are under the control of a manually operated switch 84. Normally closed limit switches 85 and 86 are provided and these are positioned so that the switch 85 is opened when the arm is fully raised or flexed and the switch 86 is open when the arm is fully lowered or straightened.
In brief, in the case of the prosthetic arm herein described, the EMG signals are derived from the biceps and triceps by electrodes held in contact with the skin of the stump overlying those muscles. Such signals are alternating and relatively weak and must be rectified and suitably amplified. The circuit sections for thus processing EMG signals derived from the biceps and triceps are indicated at 87 and 88, respectively, see FIG. 10.
The electrical energy is applied to the motor 37 in pulses with their width determining the voltage signal to the motor. The polarity of the voltage input to the motor 37 and the width of the pulse is determined by EMG signals and the pulse amplitude is determined by the battery 32.
The processed signals are combined in a summer, generally indicated at 89 and schematically detailed in FIG. 11 to provide a difference voltage orsignal representative of the strength difference between the processed input EMG signals. By way of example, the difference velocity is zero when the bicep and tricep signals are of equal magnitude. If the signal derived from the bicep is the larger, the prosthesis is flexed and if the tricep signals are the larger, a reverse or straightening movement results. Such combined signals are further processed by an amplifier 83, see FIG. 14, after modification by a pulse width modulator 90, see FIG. 12.
The tachometer 38 has its feedback connected to the summer 89 to provide a signal proportional to the rate of movement, see FIG. 12. Reference'has already been made to the use of the velocity feedback to reduce the power to the motor 37 when a heavy load is being lowered. The velocity feedback from the tachometer 38 subtracts from the processed .EMG signals and acts to overcome the effect of the non-linearities in the friction of the mechanical drive system.
Below follows a detailed description of the electronic circuitry as shown in FIG. 10 through FIG. 14. Since the signal amplification and processing is the same for both the biceps and triceps, only the biceps channel is detailed but with the corresponding components of the triceps channel distinguished by the suffix addition vA to the appropriate reference numerals. The two biceps electrodes 92 and 93 pick up the EMG signal from the biceps muscle surface from where it is conducted into amplifiers 94 and 95. These amplifiers are unity gain followers and present a very high input impedance to the electrodes. The inputs of each of the amplifiers 94 and 95 are protected from excessive voltage surges by a resistor 96 in conjunction with diodes 97 and 98 and said amplifier is stabilized against oscillation by a capacitor 99.
The output of each of the amplifiers 94 and95 feeds into an amplifier 100, the output of the amplifier 94 with a positive sign and the output of the amplifier 95 with a negative sign. Thus the amplifier 100 amplifies only the difference between these two signals and rejects any signal that is common to-both of them. The voltage gain of amplifier 100 is 10 while the common mode rejection is 60 dB. The gain of amplifier 100 is set by resistors 101, 102, 103, and 104. Means are provided to permit trimming of the resistors 101 or 102 to improve the common mode rejection beyond 60 dB. Capacitor 106 stabilizes the amplifier 100 whose output feeds through the capacitor 107 and the resistor 108 and into amplifier 109.
The amplifier 109 serves several functions. It provides a high gain of up to 5,000 and filters out undesirable frequencies. Filtering networks are the capacitors 107 and 110 and the resistors 108, 111, and 112 and the capacitors 113, 114, and 115. Maximum gain of this amplifier stage is determined by the resistors 108, 116, 117, I18, and 111 and the resistor 118 is adjustable to permit some gain control. The output of the amplifier 109 is an AC signal proportional to the biceps EMG signal but amplified about 40,000 times. The same processing occurs in the triceps EMG amplifier and the amplified triceps EMG signal appears at the output of amplifier 109A.
Rectification of these two signals is accomplished by diodes 119, 120, 121, and 122 in conjunction with the amplifier 123. The bicep signal is full wave rectified with a positive sign and the triceps is full wave rectified with a negative sign. Since both rectified signals feed into amplifier 123 its output produces the difference of these two signals. The output voltage is positive if the biceps EMG signal is stronger, the output is negative if the triceps EMG signal is stronger and the output is zero if both biceps and triceps signals have the same strength. Resistors 124 and 125, and capacitors 126 and 127 set the gain of amplifier 123 to 10 and provide filtering of any AC ripple remaining on the rectified signal. The capacitor 128 stabilizes the amplifier 123. The diodes 119, 120, 121, and 122 also provide some incidental deadband effect due to the nonlinear characteristics of a semi-conductor diode. The deadband discriminates against 60 Hz interference signals because it is sensitive to the peak value of voltage. 6OHZ interference signals are essentially sinusoidal in shape while EMG signals contain many voltage spikes. Thus for the same voltage average of 60 Hz and EMG the EMG will have much higher voltage spikes the tips of which would pass through the deadband while the 60 Hz sinewave amplitude would not be high enough. This and other deadbands will be described later.
FIG. 11 shows the circuitry required for the nonlinear filter, the velocity amplifier, and the summary amplifier of the summer 89.
The EMG signal contains a large amount of random amplitude variation even if the muscles producing the signals are under relatively constant tension. Since it is desirable to hold the elbow stationary for some tasks,
and fast amplitude variations improves the control of i the elbow appreciably. The filter uses the nonlinear V-I characteristics of di odes 129, 130, 131, and 132 in conjunction with a capacitor 133. An amplifier 134 is used as a unity gain follower providing very small loading for the filter. The time constant of the filter is determined by the effective resistance of the diodes which in turn depend on the diode current. For sudden large input voltage changes the diode current is high and the resulting time constant will be small. For a small and slow input voltage variation, the time constant is large. Upper and lower limits on the time costant are provided by the resistances 134 and 135. The amplifier is stabilized by' a capacitor 137. A second deadband is produced by the diodes 138 and 139. a
The velocity signal amplifier 140 receives its input signal from the tachometer38. The amplifier l40has a gain of one-half which is set by resistors 141 and 142 and a capacitor 143 stabilizes the amplifier 140.
The summing amplifier 144 of the summer 89 combines the processed EMG signal and the velocity signal in the proper proportion as determined by the resistors 145 and 146 and a resistor 147 sets the gain of the am plifier 144 and the capacitor 148 stabilizes it.
The power amplifier 83 receives the signal from the summing amplifier 144 and drives the torque motor 37 proportionately. To minimize power losses in the power amplifier 90, it operates in a switching mode. Its outputs are voltage pulses of constant amplitude and repetition rate but varying in width according to the input signal. The amplifier 83 can be separated into sections having functions: (1) the pulse width modulator 90A, see FIG. 12, (2) the polarity sensor, (3) the power switching amplifier, the circuitry shown in FIG. 14.
FIG. 12 shows the circuitry necessary to convert the output of the summing amplifier 144 of FIG. 11 into the pulse width modulated signal required to drive the motor 37. To simplify the switching amplifier and pulse width modulator design, the output from summing amplifier 144 is split into an unidirectional amplitude signal and a polarity signal. The amplitude is pulse width modulated, while thepolarity signal determines the direction of current through the motor. Both of these functions are accomplished with the amplifier 145. A-
unidirectional amplitude signal is obtained by rectifying the incoming signal from the summing amplifier 144 output. The rectification is obtained by the diode 146 and the use of the base-emitter diode of a transistor 147. These two diodes create a half wave rectified signal at the junction of the diode 146 and a resistor 148. Full wave rectification of the signal occurs at the junction of the resistors 149 and 150. Resistors 151 and 148 set the gain of amplifier and the capacitor 152 stabilizes amplifier 145.
The polarity signal is generated by transistor 147. When the base-emitter diode of the transistor 147 conducts, i.e., when the input to the amplifier 145 is positive a current will flow in the resistor 153, turning on the transistor 154 and causing signal P+ to go positive. At the same time signal P- will go towards ground. When'the input signal to the amplifier 145 is negative the current through the base-emitter diode of the transistor 147 will go to zero and will cut off, letting point P- rise toward +5V and causing P+ to fall to ground potential. A resistor 155 provides a path for the base leakage current of the transistor 154 while the resistor 156 acts as the collector resistor.
Pulse width modulation is accomplished by the amplifier 157 which compares the rectified output of the amplifier 145 with a triangular wave generated by the amplifiers 158 and 159 When the triangular signal exceeds the rectified signal the output of the amplifier 157 will be negative. When the. rectified signal exceeds the triangular wave the output of the amplifier 157 .Will be positive. The output of the amplifier 157 will thus be a square wave of constant amplitude, with the same frequency as the triangular wave and a pulse width proportional to the amplitude of the rectified signal.
.The triangular signal is generated by the amplifier 158 which is an integrator and by the amplifier 159 which is a comparator, the former using a resistor 160 and a capacitor 161 to generate a ramp whose slope depends on the polarity of the current flowing through the resistor'160. When the voltage at the output .of the comparator amplifier 159 is positive, current will flow into terminal No. -2 of the amplifier 157 and the ramp will have a negative slope. The output of the amplifier 157 is compared against a signal derived from the output of the amplifier 159 through resistors 162, 163, and 164 and diode 165. When the ramp signal falls below the reference voltage level established by the resistors 163 and 164 and the diode 165, the comparator 159 will switch to a negative and cause a negative reference voltage level and also reverse the direction of the current in the resistor 160. The triangular signal excursion extends from0 volts to some positive level with its zero level determined by the resistors 163 and 164 and the diode 165. This switches the integrated signal from a negative to a positive slope with the output of the amplifier 158 rising. When the output of the amplifier 158 rising. When the output of the amplifier 158 overcomes the reference voltage the output of the amplifier 159 will switchpositive again and the cycle will repeat. Thus the output of amplifier 15 becomes a triangular signal. A resistor 166 provides a matching input impedance for the terminal No. 3 of the amplifier 159 while a capacitor 167 stabilizes the amplifier. The resistor 163 establishes a negative bias for the reference voltage of the amplifier 159. This creates a deadband for the pulse width modulator. Resistors 168 and 169 adjust the amplituce of the triangular signal to its proper level.
In FIG. 13, a battery filter is shown having resistors 170 and 171 and capacitors 172 and 173 to filter the battery voltage to prevent noise from entering the operational amplifier circuits.
The switching amplifier 83 shown in more detail in FIG. 14 must serve two functions. It must control the total power applied to the motor 37, i.e., the elbow torque and it must also control the polarity of this power, i.e., the direction of the torque. Power flowing into the motor 37 is controlled by four power transistors which operate in pairs. When the transistors 174 and 175 are conducting the motor 37 will drive the elbow unit 33 in a flexing direction. When the transistors 176 and 177 are turned on, the motor 37 will drive the unit 33 to extend or straighten the forearm 19. A flip-flop comprised of transistors 175 and 178 decides which power transistor pair controls the motor power. Action of the flip-flop will be described later.
The amplitude of the motor current is controlled by the on-time of the power transistors which again is a function of the pulse width modulator signal derived from the output of the amplifier 157, see FIG. 12, and is applied through the resistor 180 to the base of the transistor 181. When this signal is positive, the transistor will turn on pulling resistors 182 and 183 to ground. This will turn on the transistor 184 or the transistor 185 depending on the state of the flip-flop. Assume that the transistor 184 turns on, the current will flow through the resistor 186 and the transistor 187 will turn on to turn on the transistor 187 directly and the transistor 174 through the resistor 188. Resistors 189, 190, 191, 192, 193, 194, and 195 are base leakage resistors. Resistors 186, 196, 197, and 188 limit the current flow. Diodes 198, 199, 200, and 201 are flyback diodes which provide for continuity of the current through the motor inductance during the switching interval thus protecting the power transistors from high voltage spikes.
The flip-flop, the transistors 175 and 178, is clocked or toggled by the leading edge of the pulse width modulator signal and its direction of toggle is determined by the state of the polarity signals P+ and P. If P+ is positive and P- is at ground potential every clock pulse will turn off transistor 175 which will turn on the transistor 178. If this was the state of the flip-flop before the clock pulse arrived no action will take place. If the transistor 17 was turned on and the transistor 179 was turned off before the clock pulse arrived, the clock pulse will toggle the flip-flop into its new state. Reversal of the polarity signals will cause the opposite behavior. Resistors 202 and 203 provide base leakage paths. Resistors 204 and 205 are the collector resistors for the transistors 175 and 178, respectively. Resistors 206 and 207 provide the cross-coupling between the transistors 175 and 178. The gating and trigger circuits uitlize resistors 208, 209, diodes 210, 211, 212, and 213, and
Necessary to the objective of providing a jointed prosthesis of appropriate weight and weight distribution is a reduction in the weight of the battery pack and the recognition of the concept that, rather than to try to duplicate the action of a natural arm, the effort should be made to give the amputee the greatest functional capability with the least inconvenience.
Battery size and weight are determined by work requirements, for example, the maximum weight that is to be lifted in a predetermined time. The use of a purely mechanical, reverse locking clutch instead of an electro-mechanical brake cut in half the quiescent drain of the system and in combination with the improved circuitry, reduced the final quiescent drain to eleven milliamperes from an original 180 milliamperes. As a consequence, the standby operating time increased from 6 to 45 hours despite a reduction in the weight of the battery source from 3 /2 pounds to about 8 ounces, the maximum lift at 12 inches to be 5 pounds, and the speed between full flexion and full extension (without load) to be 1.0 second.
Extraneous EMG signals are, however, produced unconsciously by the amputee and these can easily produce momentary current drains in excess of 350 milliamperes making it essential to introduce a deadband into the signal processing system. Without such a deadband, current is delivered to the motor as a result of such extraneous EMG signals although the motor does not run as the motor does not produce enough torque to overcome starting friction. The object of the deadband is to provide, in brief, that zero effort provides zero current to the motor.
While the diodes 119, 120, 121, and 122 of the amplifier 123 incidentally delete some of the extraneous EMG signals an additional deadband is necessary for the effective control of such extraneous EMG signals. An effective deadband is, accordingly, provided before the summer 144 by the parallel diodes 138 and 139.
A problem requiring an additional or second deadband is that DC drift and noise create signals from the junction where EMG and velocity feedback signals are summed by the amplifier 144. This additional deadband is found in the pulse modulator stage. The triangular signal excursion in the generation of triangular waves extends from aboutO volts to some positive level and its lower point is determined by the combination of the resistors 163 and 164 and the diode 165. The resistor 163 biases the zero point to a positive value to eliminate the lower 10 percent of the summing voltage thereby providing a deadband to prevent small variations that might occur about the zero level.
The assignment of proper values of the second deadband is difficult due to conflicting requirements within the feedback loop of the system. On the one hand, it is desirable to make the deadband as small as possible to minimize its effect on the control of the arm. At the same time, the maximum allowable plant gain is directly determined by the magnitude of the deadband with a larger gain value resulting in the system being stiff with respect to velocity which is desirable as it eliminates the effect of stick-friction in the mechanical drive system.
A large deadband is desirable for good servo system characteristics. The relationship between the deadband value and the controlability of the arm is more difficult and is best determined in relationship to the mechanical deadband created by thefriction in the system and if less than that, say by one-half, the amputee will feel it as a slight increase in the mechanical deadband without noticeable loss of control. In practice, each deadband is in the neighborhood of percent of the maximum signal levels.
It will thus be appreciated that prosthetic joints in accordance with the invention are well adapted to meet the varied requirements involved in providing maximum functional capability with the least inconvenience to the amputee.
1. A prosthesis comprising a first member having a stump receiving socket and provided with a support at its distal end, a second member, a joint unit secured to the support of the first member, a pivotal connection between saidjoint unit and the second member relative to said first member, a drive including a direct current permanent magnet motor and a transmission housed within the unit with the output shaft of the transmission so connected tosaid second member that said second member is flexed relative to said first member in a direction dependent on the direction in which said motor is operating, said transmission including first and second reduction stages and a reverse locking clutch between said stages operable under the conditions that the second member is in support of a load and the motor is not operating to move the member to raise or lower the load and electric circuitry including said motor, a battery, electrodes attachable to the stump and operable to detect EMG signals from the stump muscles that were used to flex and straighten the amputees missing joint, and signal processing means operable to convert the detected signals into signals that operate the motor in the direction and rate that corresponds to the strength of the dominant EMG signals.
2. The prosthesis of claim 1 in which the first stage is a planetary gear reduction unit and the second stage is a plano-centric drive.
3. The prosthesis of claim 1 in which the transmission also includes a friction brake that may slip to provide a cushion and the engagement of the reverse locking clutch.
4. The prosthesis of claim 1 and a direct current permanent magnet tachometer coupled to the motor and operable to provide a signal proportional to the rate at which said second member moves relative to the first member, the signal processing means includes means to amplify and convert the signals indicative of flexing to one polarity and those indicative of straightening to the other polarity, a summer operable to combine said signals to provide a difference signal, the tachometer provides a feedback signal proportional to the rate of flexing velocity, means summing the feedback and the difference signals to decrease the power input when said rate is greater than that intended by the wearer of the prosthesis, and a pulse width modulation and motor driving amplifier in which battery circuit determines the amplitude of the pulses.
5. The prosthesis of claim 4 in which the reverse locking clutch includes a friction brake operable, when a load is being lowered to cushion by slipping the reengagement of the clutch as the motor and the load alternatively release and lock the clutch.
6. The prosthesis of claim 1 in which the circuitry includes deadband means operable to exclude substantially all extraneous signals that would, when zero movement of the second member was wanted, cause battery drain.
7. The prosthesis of claim 6 in which the value of the deadband means is less than the mechaical deadband created by friction of the unit.
8. The prosthesis of claim 1 in which the circuitry includes limit switches operable at either of the limits to which the second member may swing to open said circuitry.
9. The prosthesis of claim 1 in which the circuitry including its power source is housed within the second member.
10. The prosthesis of claim 1 in which the first stage of the transmission includes an internally toothed ring gear, a gear fast on the motor drive shaft, and a rotatable member includes planetary gears in mesh with said ring and drive shaft gears, and the clutch includes a housing and a driven member, means connecting the housing and the ring gear to the motor casing and a driving connection between the rotatable member and the driven member of the clutch.
11. The prosthesis of claim 10 in which the driven member of the clutch includes an output cam having oppositely disposed parallel flats and a shaft, the rotatable member includes oppositely disposed spaced pairs of connectors, the pairs receiving said flats freely between them to provide said driving connection when the motor is in operation, and rollers confined freely between the connectors of each pair but permitting them to become jammed between the clutch housing and the flats when said second member is moved in one direction when the motor is not driving said second member in that direction.
12. The prosthesis of claim 11 in which said one direction is the joint-straightening direction.
13. The prosthesis of claim 11 in which the clutch housing includes a flange and an outer clutch housing clamps said flange against the face of the ring gear to an extent providing frictional resistance against the flange turning in either direction.
14. The prosthesis of claim 11 in which the second stage is a transmission of the plano-centric type with its driven shaft the output shaft of the unit, the second stage including a fixed circular spline, the motor, the first stage including a fixed circular spline, the motor, the first stage and the clutch are housed within the second stage and the motor is connected thereto to rotate therewith. and the fixed circular spline of the second stage is fixed within the unit. 15. The prosthesis of claim 11 and a DC PM tachometer coupled to the motor and operable to provide a signal proportional to the rate of flexing velocity so that when the rotatable member unjams the rollers and the external load on the arm causes the output cam to go in the same direction the feedback signal prevents the arm from taking off in a mannor not intended by the amputee.
16. The prosthesis of claim 11 in which the acute angle which a line perpendicular to the face of each flat of the output cam and which passes through the center of the appropriate roller would make with a second line also passing through the center of the roller and through the point of its tangency to the inside of the housing is no less than l2.14 and no more than 15. 14 in order to minimize the force required by the rotatable member to unjam the rollers.
17. The prosthesis of claim 4 and a deadband following the signal-combining summer providing the difference signals to bar extraneous EMG signals, said deadband being in the neighborhood of ten percent of the maximum signal level.
18. The prosthesis of claim 17 in which the deadband motor driving amplifier, a triangular wave generator to provide the amplitude of the pulses and having a signal excursion extending from 0 volts to a positive level,
said zero level comprising first and second resistors and a diode, said first resistor comprising the second deadband and biasing said zero point to a positive value that eliminates voltage in the neighborhood of ten percent of the voltage output of said generator.
21. The prosthesis of claim 4 in which the pulse modulating and motor driving amplifier includes two pairs of power transistors, each pair for a particular polarity of the pulse delivered thereto, and a flip-flop operable to select the appropriate pair of power transistors that is to be operated to control direction in response to the polarity of the polarity signals and shiftable only on and at the leading edge of the pulse width modulation pulse.