Note: Descriptions are shown in the official language in which they were submitted.
1~3~77S4
Background of the Invention
It is important to realize that in the operation of a laser
seeker, two types of signal level variations exist. One of these is
the variations of the average input signal level with range, and the
other is a change in instantaneous signal level due to scintillation,
foreground objects, etc.
Laser Seeker signal processors conventionally employed for
proportional tracking utilize linear signal amplification of a type which
limits the total instantaneous dynamic range to approximately 20 db, or -
10 db about the average pulse amplitude. However, the scintillation in
the reflected laser energy from a target caused by missile and illumlnator
aiming motion can cause pulse to pulse amplitude variations exceeding
20 db. The resulting saturation or dropping of pulses will reduce the data
rate and degrade guidance accuracy. In addition, terrain masking can
occur, which is responsible for creating false pulses and preventing
a large percentage of the energy from reaching the target.
In several instances, during field tests of laser illumLnated
tactical tArgets, a 25 db variation from pulse to pulse was observed due to
scintillation and terrain masking. Under these conditions the 20 db
instantaneous dynamic range of a conventional proportional processor will
cause pulses to be lost with the resultant degradation in accuracy.
Further, with conventional processing equipment, it is possible
to get a series of returns, such as from foreground bushes or other objects,
which will have the effect in signal processors of limited dynamic range of
causing the signal processor to track the false return. It will be seen that
l~q~75~
if a false pulse arrives earlier than the true pulse and is of
a higher amplitude, if the false pulse is greater than 1/2 of
the instantaneous dynamic range than the true pulse, then the
system may lock upon the false target. If this type of
situation is to be avoided, the signal processor must have a
wide instantaneous dynamic range.
Summary of the Invention
According to the invention there is provided a method
of processing two detector output signals related to a common
sensing plane so as to produce a normalized output signal
related to the two detector output signals, comprising applying
the detector output signals to separate signal channels each
containing a logarithmic amplifier in order to produce signals
representing the logarithms of the detector output signals;
converting signal outputs of nanosecond duration from each
logarithmic amplifier into signals of at least sufficient
duration to provide sufficient time for comparison of the
channel output signals thereby produced; comparing the channel
output signals to produce a steering command output whose
polarity is indicative of the channel having the higher output
signal level and which has a slope representing volts versus
target bearing which is independent of the input signal level
thus inherently accomplishing normalization; and limiting the
steering command output to a preselected voltage level below
a non-linear region of that output.
The invention also provides a proportional signal
processor for processing two detector output signals related
to a common sensing plane by the method of the invention, comprising
3 --
.,
10"7754
detector means for sensing the bearing of a target; at least
one pair of channels connected to the detector means to receive
detector output signals related to a common sensing plane
which output signals represent target position; a logarithmic
amplifier in each signal channel each connected to receive
one of the detector output signals and to produce a signal
related to the logarithm of a wide range of levels of the
received detector output signal; sample and hold circuit
means in each signal channel for converting pulse type signal
outputs of nanoseconds duration into signals of at least
sufficient duration to provide sufficient time for comparison of
the signals in the two channels; difference amplifier means
connected to receive the outputs of said channels and operable
to compare the outputs and produce a steering command output
whose polarity is indicative of the channel having the higher
output signal level and which has a slope representing volts
versus target bearing which is independent of the detector
output signal level thus inherently accomplishing normalization;
and limiter means for limiting the output of said difference
amplifier means to a preselected voltage level below a non-
linear region of that output.
The invention thus provides a signal processor having
an instantaneous dynamic range of - 30 db, which enables my
system to distinguish true target returns from false ones.
Whereas previous proportional type signal processors were of
limited capability, the present invention makes possible an
instantaneous dynamic range of 60 db or greater in a proportional
system, and accurate tracking information will be provided when
- 3a -
10~7754
pulse to pulse variations are as great as + 30 db from the
average signal level. The instantaneous dynamic range of my
invention is to a large extent determined by the dynamic range
capability of a logarithmic amplifier utilized in the
amplification arrangement of each channel of my device. However,
even a logarithmic amplifier may not by itself have sufficient
dynamic range to cope with signal variation due to scintillation
when this is superimposed with variation due to range closure
changes.
Accordingly, I use an AGC arrangement that enables the
amplification arrangement to operate about the middle of its
linear range, which makes it possible with a 60 db logarithmic
amplifier to handle a pulse to pulse variation equivalent to
the square root of 1,000, which is about 31.6 to 1.
- 3b -
It is to be noted that normalization previously obtained only by the
use of additional circuitry is inherently accomplished in accordance with
the present invention by taking the quotient of the logarithm of the up and
down channels, which produces a steering command voltage whose slope is
independent of signal level.
As should now be apparent, the distinct advantage of extremely
wide dynamic range is thus obtained by the utilization of the logarithmic
amplifies, while at the same time, circuit complexity and cost are reduced
inasmuch as normalization of the guidance signal is inherent in the signal
processing utilized in accordance with this invention, thereby eliminating
the discrete normalization circuits previously needed. A further cost
reduction results from the fact that matched gains or tracking in the video
amplifiers is required between only two channels rather than necessitating a
matching between four channels as required in present proportional processors.
Thus, wide instantaneous dynamic range in accordance with this invention,
in conjunction with last pulse logic, uill allow my signal processor to track
the true target even though the true target is 30 db lower than an earlier false
pulse.
It is therefore a primary object of the present invention to provide
a signal processor having extremely wide dynamic range.
It is another object of this invention to provide a signal processor
for use in conjunction with laser illuminators and the like for proportional
tracking, which provides an instantaneous dynamic range of 60 db or greater.
It is yet another object of this invention to provide a signal processor
having increased dynamic range obtained with reduced circuit complexity.
It is still another object of this invention to provide a signal processor
of reduced complexity and cost, made possible because the normalization of
the guidance signal is inherent in the signal processing technique utilized.
10~7 7~4
These and other objects, features and advantages will be more
apparent from a study of the drawings in which:
Figure 1 is a somewhat simplified block diagram illustrating my
novel signal processor in conjunction with a more or less conventional quad-
rant detector;
Figure 2 is a graph which shows the steering command voltage
produced by the present invention as a function of target motion about the
optical system boresight axis; and
Figure 3 is a graph which shows the relation between the received
target energy level and the sample and hDld output level as controlled by the
AGC system.
Detailed Descrip ion
T~rning now to Figure 1, I have there revealed a typical e~bodi-
ment of my signal processor 10, which is shown operatively associated with
a detector 12. The detector is disposed in such a position that light from,
for example, a laser illuminated target 14 is imaged thereon as a defocused
spot by means of a suitable optical system, represented here by a lens 16.
The detector 12 may be a four quadrant PIN diode, utilizing quadrants iden-
tified as A, B, C, & D, reading in a clockwise direction. As will be under-
stood, my signal processor may be used as an intrinsic part of a cccker head
of a missile, for example, but is obviously not to be so limited. As an
example, the detector may relate to use with a non-optical arrangement, such
as an RF direction finder in which the incoming signal is detected by four
directional antennas. Thus, an emkodiment of my invention can use either
a quadrant type detector as illustrated, or can use four related but separate
detectors.
-- 5 --
My signal processor will be explained in conjuction with a pair
of channels related to the same sensing plane, such as with the channels con-
cerned with the derivation of up-down commands. The channels shown in
Figure 1 provide a pitch guidance signal proportional to the vertical displace-
ment of the defocused spot frcm the center of the detector, and in accordance
with the teachings of the present invention, pitch error is equal to Log (A+B)
m mus Log (C+D). However, it is to be understood that the signal processor
for the orthogonally related channel is essentially identical, except that of
course the yaw error is equal to Log (A+D) minus Log (B+C). The processor
for the left-right channels therefore does not need to be separately treated
here.
It will De seen in Figure 1 that the output signals from quadrants
A, B, C and D are delivered to respective preamplifiers 18, 20, 22 and 24,
each of which has a bandwidth of 25 megacycles. Incorporated in each pre-
amplifier is a diode attenuator network that makes gain control possible. The
signal handling range of each preamplifier is 60 db, with the AGCing of the
diode attenuators in a manner described hereinafter affording an additional
90 db of gain control.
Figure 1 further reveals that the outputs of preamplifiers 18
and 20 are summed in a linear summing amplifier 26, the bandwidth of
which is 35 megacycles and the gain of which is unity. Simiarly, the
outputs of preamplifiers 22 and 24 are summed in a linear summing amplifier
28, the characteristics of which are identical to those of summing amplifier
26.
754
The outputs from summing amplifiers 26 and 28 are respectively
applied to logarithmic amplifiers 30 and 32, the gain charac eristics of which
may for example be logarithmic over a 60 db dynamic range. The utilization of
logarithmic amplifiers is of key importance to my signal processor in that
the function they provide to the circuit to a large extent makes possible the
wide dynamic range capability of my device, but the log amps per se are not
a part of my invention, and may for example be of integrated circuit construc-
tion, such as are obtainable from Texas Instruments and others. Thus, the
use herein of amplifier means including log amps makes possible the ampli-
fication of wide range of signal levels.
The output from log amps 30 and 32 is resepctively connected
to sample and hold circuits 34 and 36, where the short duration pulses, such
as 15 nanosecond pulses from a laser, may be stretched to hold a constant
value between consecutives pulses. The sample and hold circuits are prefer-
ably known devices of a two stretch type, that serve to stretch the pulses
from the nanosecond to the millisecond region in accordance with conventional
practice.
Steering ccmmands are developed by now taking the difference
of the tWD sample and hold outputs, this being accomplished by connecting
the sample and hold devices 34 and 36 to a difference amplifier 40, the
output of which is linear over an angular region of the detector 12 corresponding
to 2/3 of the radius of the defocused spot. The output ~rom the difference
amplifier is delivered to a limiter 42 in order to provide a steering ccmman~
at output 44 of constant amplitude heyond the 2/3 radius point. Figure 2
754
reveals the clamping of the signal at an appropriate location to provide a
constant amplitude steering ccmmand beyond the linear region.
The outputs of the sample and hold circuits 34 and 36 are also
applied to an OR gate 46 and thence to AGC 48 in order to develop an ACG
voltage. A hold off bias is applied to the OR gate so that the average signal
level rises 30 db above threshold (1/2 of the logarithmic range of the ampli-
fiers prior to the development of an ACG voltage). After the AGC threshold
is reached, the AGC output voltage is fed from 48 kack to the four preampli-
fiers 18, 20, 22 and 24 as shown in Figure 1 in order to maintain the output
of the sample and holds at a constant value. This AGC is effective over an
additional 90 db of dynamic range.
The AGC arrangement serves to hold the average signal strength
in the middle of the log amp dynamic range by suitably changing the gain of
the preamps. This gain change can be accomplished for example by the use
of a diode attenuator netw~rk, as previously mentioned. Typically, if a 60 db
instantaneous dynamic range is required, a 30 db threshold w~uld be employed
in the AGC. No AGC would be developed through the OR gate 46 until the
stronger of the tWD channels exceeds 30 db above threshold~ The AGC w~uld
then be applied to the linear amplifiers to maintain the average pulse amplitude
at the midpoint of the 60 db log amp dynamic range. Pulse to pulse variations
of - 30 db cound thus occur without affecting the accuracy of the proportional
tracking signal.
The instantaneous dynamic range of my design is dictated by the
ratio of the main loke to side loke energy of the target illuminator. With
present day state of the art laser illuminators, an instantaneous dynamic
1~97~S4
range of greater than - 30 db might well result in the processor tracking
fal æ target created by side lobe energy. An A~C system is therefore
highly desirable in conjunction with the log amp dynamic range, to cover the
120 db total dynamic range required by most laser _eekers. However, a
total log amplifier range of 120 db is possible with the present state of the
art, and if u æd, would elimLnate the need for the AGC arragnement.
In operation, energy reflected from the target 14 is received
through the optical system and imaged onto the four quadrant detector 12.
The signal from each quadrant is amplified in a linear manner by the respec-
tive preamplifiers 18 through 24. The signals from the A and B quadrants
are summd in the summing amplifier 26, applied to the log amplifier 30,
and the pulse output from the log amplifier representing the logarithm of the
(A + B) sum signal is then stretched for one int~r pulse period in the sample
and hold circuit 34. Similarly, the signals from the C and D quadrants
are amplified in the preamps 22 and 24, c~mbined in the summing amplifier
28, and applied to the log amplifier 32. The output of the log amplifier
repre ænting the logarithm of the (C + D) sum signal is then stretched in the
sample and hold circuit 36 for one inter pulse period.
The difference in the sample and hold outputs is then taken in
the difference amplifier 40, coupled through the limiter 42, which then pro-
duces at 44 the steering command for the up/down channel. As the missile
closes on the target, the average signal level at the output of the sample and
h~lds 34 and 36 will increase and when it reaches a point 30 db akove thres-
hold, the biased diodes in the O~ gate 46 will couple a signal to the AGC 48,
which will ~e fed back to the preamps 18 thorugh 24 in such a manner as to
~. ~
10~7754
maintain the larger of the sample and hold outputs at a constant amplitude.
Most importantly, therefore, the output of the sample and holds is held con-
stant at the midpoint of the logarithmic range of the log amps, despite range
closure.
The signal may therefore vary on a pulse-to-pulse basis by a
factor of + 30 db from the average value once the AGC threshold has keen
reached, without loss of signals. This enables, through utilization of last
pulse logic, my seeker to develop accurate guidance information, even in the
presence of scintillation that produces large pulse-to-pulse signal variations,
or even in the presence of terrain masking, which produces false signal
returns which may be as much as 30 db greater than the true target return.
Last pulse logic develops the steering information from the last
signal energy which exceeds the threshold sensitivity of the system. ~ach
pulse which exceeds the threshold is processed and stored in the sample and
hold circuits. A succeeding pulse discharges the steering informations~ored
in the sample and hold from the previous pulse, and therefore produces a
steering ccmmand from the last pulse only.
Inherent also in the processing technique in accordance with this
invention is the implementation of the steering commands so that the steering
ocmmand voltage which is proportional to the angle between the optical axis
and t~rget bearing remains constant over a range of signals of - 30 db about
the average value.
The normalization technique inherent in this implementation is
accomplished by taking the quotient of the logarithm of the up and down chan-
-- 10 --
:lOg7754
nels, which produces a steering command voltage slope is independentof signal level. Significantly, as taught herein, the numker of parts required
in order to obtain normalization is appreciably reduced from the number re-
quired in the usual normalization procedures, which necessitated taking
A + B, subtracting C + D, and then dividing the sum of A + B + C + D.
My signal processing technique is capable of use in other applica-
tions, such as in a monopulse R. F. direction finder. In an RF direction
finder each of the quadrants of the detector would be replaced by an antenna
and RF detector. The processing procedure would be very close to that shown
herein.
The proportional steering command which is produced by my sig-
nal processing technique is shown in Figure 2. This steering command
signal which is linear over a region of - 2 about the boresight axis was pro-
duced by a system using a 3 radius defocused spot. The defocused sFot
size may be varied to obtain the desired linear region. The solid curve is
the steering ccmmand produced at the output of the difference amplifier 40 as
the target is positioned so as to move the defocused spot on the detector over
a region of - 3 about the boresight axis. Figure 2 shows that the steering
command voltage is linear with angle over a region which is approximately
2/3 of the defocused sFot size, or - 2. The steering ccmmand is therefore
restricted to the linear region by means of limiter 42 which limits the differ-
ence amplifier output 40 to + 5 volts beyond the +2 angle and to -5 volts
bey~n~ the -2 angle, as shown by the dashed curve in Figure 2.
-- 11 --
1~)977~;4
A further advantage resulting from the inherent configuration of
ny signal processing technique is that the steering command signal is normal-
ized, that is the slope of the steering command voltage vs. angle off axis is
independent of the target signal strength over the full dynamic range of the
logarithmic amplifier. If the target is moved off the boresight axis by a
given amount, for example 1, the ratio of the target signal powers produced
in the A + B and C + D channels is constant and is independent of target signal
level. The difference in the logarithms of the two constant ratio pulses
(the difference of log amp 30 and 32 outputs as measured by difference ampli-
fier 40) is a constant voltage which is independent of the aboslute value of the
signal pulses. The slope of the resulting steering command 44 produced at
the output of limiter 42 is therefore independent of target signal level and a
normalized steering conmand signal is obtained with a significant simplifica-
tion in circuit complexity over the convention normalization technique.
The primary advantage of my signal processing technique is the
improved quidance ac acy resulting from a wide instantaneous dynamic
range in a proportional tracking system. Large pulse to pulse signal varia-
tions oc in both RF and optical seeker systems because of scintillation
and/or terrain masking as in the case ofa moving target illuminated by a
ground or airborne mounted laser. Variations also oc in the pulse to
pulse output from present state-of-the-art lasers. Target signature measure-
ments of tactical laser illuminated targets revealed variations approaching
- 30 db. The limited dynamic range of present generation proportional
laser ~ccks, usually -+ 10 db, will result in reduced guidance accuracy
through reduced data rate (individual pulses falling below the threshold level)
or saturated pulses (individual pulses exceeding the linear range).
- 12 -
1097754
In order to utilize the 1 30 db dynamic range offered by my signal
processing method, an AGC system must be employed which maintains the
average signal amplitude at the midpoint of the instantaneous dynamic range.
The AGC characteristic for my slgnal processing method is shown in Figure 3.
The output of the sample and hold outputs 34 and 36 rises from the threshold
value of 1 volt to 5.5 volts (+ 30 db ab~ve threshold) as the seeker closes on
the target. The OR circuit 46 bias is exceeded at a sample and hold output
voltage of 5.5 volts and the AGC 48 develops a gain control voltage which is
fed back to the diode attenuator netwDrks in the preamps ]8 through 24. The
output from the larger of the sample and hold outputs, represented by the
horizontal line in Figure 3, is held at 5.5 volts for an additional increase of
90 db. The instantaneous pulse amplitude, represented by the sloped line
in Figure 3, may therefore vary by - 30 db from the average value with~t
loss of accuracy once the average signal strength has risen 30 db above
threshold. For illustration the instantaneous load line of - 30 db is drawn
at the + 90 db signal level above threshold point in Figure 3. This instan-
taneous operating line actually progresses from the + 30 db ab~ve threshold
point toward the higher signal levels as the seeker closes on the target.
An instantaneous dynamic range of - 30 db and a total dynamic
range of 120 db are used for illustration only. The instantaneous dynamic
range and total dynamic range may be varied to meet application requirements.
As should now be apparent, I have provided a highly advantageous
logarithmic proportional signal processor admira~ly suited for use with laser
see~ers, in that it provides a significant increase in dynanic range, accom-
plished ~y circuitry whose cost and complexity are considerably reduced from
the ordinary. Because the signal outputs are the logarithm of the signal
- 13 -
1097754
inputs, the slope of the steering ccmmand is independent of input signal
amplitude, and normalization is inherent. It should further be noted that
for each factor of 10 increase in power, I obtain the same ~ in the output
voltage.
Significantly, my device thus produces, independent of the abso-
lute signal level represented by the def sed spot, a steering command
whose slope (volts vs. degrees off axis) is constant for a given angular
displacement of the defocused spot away from the midpoint of the detector,
with this being true irrespective of whether the input is n OE the threshold,
or at the upper end of the dynamic range, which of course may be a value
one million times greater.
A preferred emkcdiment of my invention may involve a signal
processor having a wide instantaneous dynamic range and usable in conjunc-
tion with a pair of channels relatable to the same sensing plane, comprising
detector means, and at least one pair of channels arranged to receive outputs
from said detector means. Amplifier means including a logarithmic ampli-
fier æ e operatively disposed in each of the channels, which are æ ranged to
receive the respective outputs of said detector means and functions to amplify
a wide range of signal levels. Difference ~l~lifier means æ e provided for
producing a signal whose polarity is indicative of the channel having the
higher output, such that appropriate commands can be generated.
Either one or two pairs of channels can be utilized, and if tw~
pairs are employed, the plane of one pair of channels is orthogonal to the
plane of the other pair of channels, thus ma~ing it p~ssible to generate
- 14 -
l~g7754
steering command usable for controlling the movement of a
vehicle, such as a missile or the like.
The signal processor may utilize automatic gain
control means, the latter means being operative for selectively
changing the gain of said amplifier means such that the
logarithmic amplifier of each channel can function substantially
at the midpoint of its operating characteristic. Sample and
hold means may also be utilized in each channel, for converting
pulse type signal outputs into signals of longer duration, thus
to provide sufficient time for the comparison by said difference
amplifier means of the outputs of said channels. Limiter means
may be provided for limiting the output of the difference
amplifier means to a preselected voltage level.
It should by now be apparent that normalization is
inherent in my invention, in that the difference amplifier
provides a steering command having a slope representing voltage
versus target bearing off boresight, that is independent of
input signal level.
The principles, preferred embodiments and mode of
operation of the present invention have been described in the
foregoing specification. The invention which is intended to be
protected herein, however, is not to be construed as limited
to the particular forms disclosed, since these are to be
regarded as illustrative rather than restrictive. Variations
and changes may be made by those skilled in the art without
departing from the spirit of the present invention.
-- 15 --
