 |
Description  |
|
|
TECHNICAL FIELD
The invention is related to position tracking systems generally and
particularly to tracking pilot helmet position in the cockpit of a flight
vehicle or simulator using ultrasonic transmitters mounted in the cockpit
and ultrasonic transducers mounted on the helmet, or vice versa.
Background Art
Position tracking systems for monitoring the position of an article such as
a pilot's helmet in a cockpit are useful for such purposes as maintaining
a computer-generated projected image display in the pilot's field of view,
for example. Other purposes include tracking any part of a person's body
such as the head, hand or finger, so that the person may generate inputs
to a computer by the movement of his head, hand or finger, for the control
of cursor movement in a computer display, for example. Other applications
include robotics and weapon systems.
The advantages of employing ultrasonic waves to perform such position
tracking over techniques employing electromagnetic signals (such as those
disclosed in U.S. Pat. No. 4,742,356 to Kuipers and U.S. Pat. No.
4,303,394 to Berke et al.) are well-known. Essentially, electromagnetic
systems are far more vulnerable to interference from many noise sources.
Ultrasonic position tracking techniques are well-known. For example, U.S.
Pat. No. 4,807,202 to Cherri et al. discloses an ultrasonic tracking
system in which the orientation and position of a movable object (such as
a helmet) in a closed frame of reference (such as a cockpit) is
continually tracked by following the six degrees of freedom of movement of
the moveable object. This is accomplished, as illustrated in FIG. 1
hereof, by mounting three ultrasonic transmitters 100, 102, 104 in three
different locations in the closed frame of reference 106 and mounting
three ultrasonic transducers 108, 110, 112 in three different locations on
the moveable object 114. Each one of the three transmitters 100, 102, 104
transmits an ultrasonic acoustic wave signal at a different ultrasonic
frequency, all of which are received at each one of the three transducers
108, 110, 112. The three ultrasonic frequencies received at each
transducer are separated into three received signals, so that a total of
nine signals are received and processed. A tracking processor 116
processes each of the nine signals to provide the distance between the
transducer and the transmitter corresponding to the frequency of the
received signal, thus providing nine distances. Using well-understood
principles, the tracking processor 116 computes the instantaneous position
and orientation of the moveable object 114 with respect to the closed
frame of reference from the nine distances. This computation uses the
locations of the transmitters in the closed frame of reference and the
locations of the transducers with respect to the frame of reference of the
moveable object.
Various methods of transmitting and processing the ultrasonic signals are
employed, all with varying degrees of limited performance. The basic
limitation of such methods is that they are slow, limiting the rate at
which the position and orientation of the movable object can be repeatedly
computed. Most of the methods employ a pulsed ranging ultrasonic
technique, such as the techniques disclosed in U.S. Pat. No. 4,853,863 to
Cohen et al., U.S. Pat. No. 4,807,202 to Cherri et al., U.S. Pat. No.
3,836,953 to Rotier and U.S. Pat. No. 3,777,305 to Stoutmeyer.
The above-referenced patent to Cohen et al. discloses an ultrasonic
position tracking technique in which the frequency shift of the ultrasonic
signal due to helmet movement is measured and integrated to provide a
displacement value from which the transmitter-to-receiver range is
computed.
The problem with the pulse or sequential techniques is that the time
required for the receiver to acquire a sufficient ultrasonic signal from
which the transmitter-to-receiver range can be inferred limits the rate at
which the helmet position can be tracked. For example, the ultrasonic
pulsed ranging techniques are limited by the time of flight between the
transmitter and the receiver. Generally, it is believed that such
sequential techniques produce an updated position measurement for a given
sensor at a rate not exceeding on the order of 10 Hz. Most users would be
pleased by a 100 Hz update rate. Pulse or sequential techniques are
limited to a total time of 9 time-of-flight intervals, or about 30 ms. In
a room with echoes, this figure may be 3 or 4 times as large.
Consequently, the maximum update rate for sequential techniques is 30 Hz,
although a more practical update rate is 10 Hz. Another problem with
sequential techniques is that the data is incoherent. The 9 ranges are not
simultaneous. As will be shown below, the present invention provides the 9
ranges simultaneously as coherent data, a significant advantage over the
prior art.
The problem with the technique of integrating the frequency shift is that
its maximum update rate is limited by the time required to measure
frequency, compute the doppler shift and then integrate it so as to
produce a displacement value from which to compute a range.
The use of plural ultrasonic frequencies to provide separate ultrasonic
signal processing at each sensor has the drawback of creating a relatively
noisy environment, which tends to detract from the system performance. An
example of an acoustical tracking system responding to many different
frequencies is described in U.S. Pat. No. 4,333,170 to Mathews et al.,
wherein there is no active acoustic transmitter, and the sound emanates
from a passive target at an unknown location. The present invention
concerns a system employing an active ultrasonic transmitter transmitting
at a single predetermined ultrasonic frequency. Unlike the Mathews et al.
system, the present invention concerns enhancing the type of system in
which sound is actively transmitted at known frequencies from fixed
predetermined locations while the sensor array is mounted on a helmet (or
other moving platform) whose location and orientation is to be determined
and updated at rapid intervals. The Mathews et al. system cannot detect
rotations of its passive target involving no translation of the target,
and therefore is unsuitable in helmet tracking systems.
What is needed is an ultrasonic position tracking system, employing a
single active ultrasonic source, which will track the moveable object
position at a high rate without requiring many different ultrasonic
frequencies.
SUMMARY OF THE INVENTION
An ultrasonic tracking system embodying the invention has an array of
plural ultrasonic detectors at discrete locations, the plural detectors
being responsive to an ultrasonic signal of frequency f emanating from a
single remote transmitter to produce respective plural electronic signals
corresponding to the ultrasonic signal as received at respective ones of
the plural detectors, and a processor for determining from the plural
electronic signals a direction of travel of the ultrasonic signal relative
to the array of ultrasonic detectors. The processor stores a matrix K
which is a function of a displacement matrix D including vectors defining
locations of respective ones of the plural detectors in the frame of
reference. The processor includes digital signal processors which sense
the phase angles of each of the electronic signals and to produce a phase
difference vector including, respectively, differences between the phase
angle of the electronic signal of one of the plural detectors and the
phase angles of the electronic signals of respective remaining ones of the
plural detectors. The processor produces an orientation vector defining an
orientation of the transmitter relative to the frame of reference of the
array of detectors by multiplying the phase difference vector by the
matrix K. The processor computes an azimuth angle and an elevation angle
from the orientation vector.
In order to provide range, a phase shift key encoder connected to the
transmitter encodes every N cycles of the ultrasonic signal with one of a
succession of encoder counts, while a phase shift key decoder coupled to
receive an electronic signal from one of the plural detectors produces a
new decoder count every N cycles of the electronic signal that are
decodable. A range processor connected to the phase shift key encoder and
to the phase shift key decoder receives simultaneously the encoder counts
and the decoder counts. From the encoder and decoder counts, the range
processor computes distance between the transmitter and the one detector.
The computation of the range is carried out by storing a first encoder
count from the phase shift key encoder, then, upon receipt from the phase
shift key decoder of a decoder count equal to the first encoder count,
sensing the concurrent encoder count from the phase shift key encoder. The
processor subtracts the first encoder count from the concurrent encoder
counts and computes the range from the resulting difference. The range is
computed by multiplying the difference by Nc/f, where N is the number of
cycles of the ultrasonic signal encoded by the phase shift key encoder
with an encoder count, c is the speed of propagation of the ultrasonic
signal from the transmitter to the array of ultrasonic detectors and f is
the frequency of the ultrasonic signal. The encoder count and the decoder
count each lie in a range of binary values between 0 and 2.sup.N -1. In
one embodiment, the phase shift key encoder employs a phase shift key code
in which respective ones of the N cycles correspond to respective binary
bits and wherein each one of the N cycles represents a binary one if
shifted by a predetermined phase shift and represents a binary zero if not
shifted. The predetermined phase shift may be 180 degrees, as one example.
The ultrasonic tracking system is preferably a digital system in which the
processor includes respective analog to digital converters having analog
inputs connected to respective ones of the plural detectors and having
digital outputs, respective digital signal processors having digital
inputs connected to the digital outputs of respective analog to digital
converters, the digital signal processors being programmed to compute a
phase angle of the signal received at its input.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a simplified diagram of an ultrasonic position tracking system of
the prior art employing three transmitters operating at three different
frequencies and three transducers.
FIG. 2 is a schematic block diagram illustrating one channel of a circuit
in the system of the parent application.
FIG.'S 3A through 3C are contemporaneous diagrams of time domain waveforms
generated in the operation of the circuit of FIG. 2.
FIG.'S 3D through 3G are diagrams of correlative phase domain waveforms
generated in the operation of the circuit of FIG. 2.
FIG. 4 is a schematic block diagram illustrating a digital implementation
of three channels of the type illustrated in FIG. 2.
FIG. 5 is a simplified block diagram illustrating the application of the
circuit of FIG. 2 with an inertial measurement unit.
FIG. 6 is a schematic block diagram of an ultrasonic tracking system of the
present invention.
FIGS. 7 and 8 depict coordinate systems of ultrasonic sensor arrays having
three and four sensors, respectively.
FIG. 9 is a diagram of a local sensor coordinate system employed in
defining a displacement matrix of the sensor array.
FIGS. 10A, 10B, 10C and 10D are time domain waveforms of various PSK
encoded ultrasonic signals representing binary values of 0, 1, 2 and 3,
respectively.
FIG. 11 is a graph illustrating the output of the PSK encoder as a function
of time.
FIG. 12 is a block diagram illustrating an algorithm carried out by a range
processor in the system of FIG. 6 to compute range.
DETAILED DESCRIPTION OF THE INVENTION SYSTEM DISCLOSED IN THE PARENT
APPLICATION
FIG. 2 illustrates a typical circuit of the invention used in each one of
the nine channels in the tracking processor 116 of FIG. 1. In the example
illustrated in FIG. 2, the circuit processes the one signal received by
the transducer 110 from the transmitter 100, it being understood that
there are a total of nine such processing tasks to be performed,
corresponding to the nine received signals discussed above with reference
to FIG. 1.
The main advantage of the invention is that it exploits the C.W. form of
the received ultrasonic signal by an interferometric comparison of the
transmitted and received versions of a signal in which any change by a
single fringe of the resulting interference pattern produces a change in
the computed tracking position. The fastest rate at which this change can
occur is the frequency of the ultrasonic signal, or about 40 KHz in one
implementation of the invention. This is an increase in tracking rate of
about two orders of magnitude over the prior art.
Overview
The interferometric displacement detector circuit of FIG. 2 includes a
fringe processor 200 having inputs connected to receive both the
transmitted and received signals, a fringe count latch 202 connected to
one output of the fringe processor 200 and a fringe fraction latch 204
connected to another output of the fringe processor 200. A master timing
generator 206 synchronizes the operation of the circuit of FIG. 2 and the
generation of the ultrasonic signal (on the order of 40 kHz, for example)
produced by the transmitter 100. The pair of latches 202 and 204 are
periodically sampled by a computer 208. As will be described below herein,
the computer 208 samples a total of nine such pairs of latches
corresponding to the nine received ultrasonic signals discussed
hereinabove. The computer 208 is programmed in accordance with the prior
art techniques referred to previously herein to compute the position and
orientation of the helmet or tracked object 114.
Accompanying the interferometric detector circuit is a rough range
processor 210 which responds to a low frequency (480 Hz, for example)
signal modulating the ultrasonic carrier transmitted by the transmitter
100 to provide a rough but unambiguous position of the helmet 114 within
the cockpit or frame of reference 106. The output of the rough range
processor 210 is held in a rough range latch 212 which is sampled by the
computer 208 as desired. The wavelength of the rough range modulation
signal must be at least on the order of the diameter or greatest length of
the cockpit or closed frame of reference 106, and so its frequency is
typically on the order of 480 Hz.
Fringe Processor
The fringe processor 200 detects the phase differences between the received
signal of the transducer 110 and the transmitted signal in quadrature pair
and deduces therefrom each zero-crossing of the phase difference. The
fringe processor 200 includes a phase detector 214 which detects the phase
difference between two signals: (a) the signal transmitted by the
transmitter 100 and signal phase-shifted by 90 degrees by a phase shifter
216 and (b) the signal received by the receiver 110 and bandpass filtered
and converted to square wave by a phase-locked loop 218. (For this
purpose, the phase-locked loop 218 is of the type well-known in the prior
art having a square wave oscillator.) Another phase detector 220 detects
the phase difference between the transmitted signal and the output of the
phase-locked loop 218. The output signal amplitudes of the detectors 214,
220 are proportional to the phase differences of the signal pairs received
by the phase detectors. The output signals of the pair of phase detectors
214, 220 are low pass filtered by a pair of filters 222, 224 and are
compared with a threshold by a pair of Schmitt triggers 226, 228,
respectively.
Operation of the Fringe Processor
FIG. 3A illustrates the time domain pulsed waveform (a square wave) of the
signal furnished by the master timing generator 206 to the transmitter
100. The reactive elements of the transmitter 100 (or, alternatively, an
active filter not shown in the drawings) convert the square waveform to a
corresponding sine wave shown in FIG. 3A, so that a sine wave C. W. signal
is received by the transducer 110. FIG. 3B illustrates the waveform at the
output of the phase shifter 216. FIG. 3C illustrates the signal received
by the transducer 110 as converted to a square wave at the output of the
phase-locked loop 218. FIG.'S 3D and 3E illustrate in the phase domain the
magnitude of the signal at the output of the low pass filters 222 and 224,
respectively, as a function of increasing phase difference detected by the
phase detectors 214, 220, respectively.
FIG.'S 3F and 3G illustrate the outputs of the Schmitt triggers 226, 228,
respectively. Essentially, each time the input to the Schmitt trigger 226
or 228 falls below its trigger threshold (dashed line of FIG. 3D or 3E),
the Schmitt trigger output goes high, and each time the input rises above
the trigger threshold the output goes low. Note that the leading edges of
the output of the Schmitt trigger 228 (FIG. 3F) coincide with the zero
phase crossings of the phase detector 220 (FIG. 3D), which is the only one
of the two phase detectors whose input is not phase shifted. These leading
edges of the Schmitt trigger 228 are detected by employing the output of
the other Schmitt trigger 226 as a gate: whenever the output of the other
Schmitt trigger 226 is low, sampling of the output of the Schmitt trigger
228 is enabled. For this purpose, the fringe counter 230 has its clock
input connected to the Schmitt trigger 228 and its enable input connected
to the other Schmitt trigger 226. The fringe counter 230 is of the
well-known type in which up-going transitions at its clock input cause the
fringe counter 230 to count up while down-going transitions at its clock
input cause the fringe counter 230 to count down. The contents of the
fringe counter 230 indicate the number of fringes or zero crossings caused
by displacement of the helmet 114.
Fringe Interpolation
The master timing generator 206 produces fringe interpolation clock pulses
at a rate which is an integral multiple (e.g., 4, 8, or 16 times) the
frequency of the transmitted ultrasonic signal. A fringe interpolation
counter 240 counts the fringe interpolation clock pulses. The rate of the
fringe interpolation clock pulses (e.g., n times the transmitted
ultrasonic signal frequency) and the modulus of the counter 240 (e.g., n)
are selected so that the counter 240 automatically resets at the end of
each sinusoidal period of the transmitted ultrasonic signal. The fringe
fraction latch 204 latches the output of the fringe interpolation counter
240 in synchronism with the received ultrasonic signal at the output of
the phase-locked loop 218. Thus, the change in phase (range) between
fringes is measured to within a fraction (1/n) of the wavelength of the
ultrasonic signal.
In the example of FIG. 2, the master timing generator 206 derives the
ultrasonic carrier for the transmitter 100 by dividing down the fringe
interpolation counter by some integer (e.g., 16). However, the ultrasonic
carrier may be derived by simply connecting the transmitter 100 to the
most significant bit of the fringe counter 240.
Rough Ranging
The rough range processor 210 operates simply by measuring the phase of the
low frequency (e.g., 480 Hz) modulation signal imposed on the ultrasonic
carrier. Since the wave length of the low frequency modulation is at least
as great as the longest dimension of the closed frame of reference or
cockpit 106, each slice is a unique location range in the cockpit 106.
As illustrated in FIG. 2, a low frequency (e.g, 480 Hz) oscillator 250
modulates the ultrasonic signal transmitted by the transmitter 100. A
conventional AM demodulator 252 in the rough range processor 210
demodulates the low frequency (e.g., 480 Hz) signal received at the
receiver 110. In the meantime, a rough range counter counts 256 up in
synchronism with the low frequency oscillator 250 and is latched in the
latch 212 by the output of the demodulator 252. Preferably, the modulus of
the rough range counter 256 and rate at which it counts are selected to
that the counter 256 resets at the end of each period of the low frequency
modulation signal. Thus, the contents of the latch 212 indicates the phase
of the modulation, which is a direct measure of the range between the
transmitter and receiver.
Multi-channel Layout
FIG. 4 is a diagram corresponding to FIG. 2 but showing at a higher level
how the three ultrasonic frequencies detected by the receiver 110 are
processed. In one implementation of the invention, the three ultrasonic
frequencies are 37 kHz, 40 kHz and 43 kHz. Thus, in FIG. 4 there are three
fringe processors 200a, 200b and 200c corresponding to the fringe
processor 200 of FIG. 2, three ultrasonic received carrier signals RC1,
RC2 and RC3 and three rough range received signals RR1, RR2 and RR3. In
addition, an address decoder 260 permits the computer 208 to selectively
address any one or portions of the various latches 204, 204 and 212, in
accordance with well-known techniques, as required by the computer's
program. The apparatus of FIG. 4 is associated with one of the
transducers, there being three transducers in all, so that FIG. 4
illustrates about one third of the system of the invention comprising a
complete tracking processor of the type illustrated in FIG. 1.
Computer
The contents of the latches 202 and 204 provide the range between the
transmitter 100 and the transducer 110 in the form of a wavelength count
(the latch 202) and a fraction of a wavelength (the latch 204). The
computer 208 is programmed in accordance with the techniques of the prior
art discussed hereinabove to periodically sample the contents of the
latches 202 and 204 for each of the nine received ultrasonic signals and
compute therefrom the helmet position and orientation.
Combination with an Inertial Measurement Unit
Referring to FIG. 5, the system 500 of FIG. 1 implemented with three
circuits of FIG. 4 for each of the three transducers 108, 110, 112 is used
to periodically update an inertial measurement unit 502. The inertial
measurement unit constantly provides position and orientation data, while
the ultrasonic position tracking system of the invention 500 provides
measurements only at periodic intervals corresponding to the position
tracking rate discussed above. However, while the inertial measurement
unit is therefore superior in its ability to continuously provide such
measurements, it suffers from a tendency to drift. The advantage of the
embodiment of FIG. 5 is that drift errors are minimized by periodically
updating the inertial measurement unit 502 with its correct position and
orientation as measured by the ultrasonic tracking system 500 at a very
high rate. For this purpose, the three transducers 108, 110, 112 are
placed on the inertial measurement unit 502, while the output of the
computer 208 embodying the tracking processor 116 of FIG. 1 is transmitted
to the inertial measurement unit 502. The inertial measurement unit
updates its position and orientation data with the measurements received
from the ult | | |
