 |
Claims  |
|
|
We claim:
1. A hybrid analog-digital pulse compressor, comprising:
an analog intermediate frequency filter receiving echo signals of a
transmitted pulse and having a passband less than a frequency sweep of
said transmitted pulse, said analog intermediate frequency filter
filtering and weighting said echo signals, and said transmitted pulse
being an FM signal;
converting means for converting output of said analog intermediate
frequency filter into digital baseband signals; and
a digital correlator for digitally correlating said digital baseband
signals to complete pulse compression of said echo signals.
2. The pulse compressor of claim 1, wherein, said analog intermediate
frequency filter is a Bessel filter.
3. The pulse compressor of claim 1, wherein said digital correlator is a
digital tapped delay line.
4. The pulse compressor of claim 3, wherein tap weights of taps in said
tapped delay line are determined by taking the inverse fast Fourier
transform of frequency response H.sub.D (.omega.) for the digital
correlator where H.sub.D (.omega.) is determined according to the
following expression:
##EQU7##
where H.sub.M (.omega.) is a desired frequency response of said pulse
compressor to achieve a specified sidelobe level and H.sub.F (.omega.) is
a frequency response of said analog intermediate frequency filter.
5. The pulse compressor of claim 3, wherein tap weights of taps in said tap
delay line are determined based on a frequency response of said analog
intermediate frequency filter.
6. The pulse compressor of claim 1, wherein said digital correlator
comprises:
a first FFT circuit taking a fast Fourier transform of said digital
baseband signals;
a second FFT circuit taking a fast Fourier transform of tap weights for
said digital correlator;
a multiplier multiplying output of said first FFT circuit and said second
FFT circuit; and
an inverse FFT circuit taking an inverse fast Fourier transform of output
from said multiplier.
7. The pulse compressor of claim 6, wherein said tap weights are determined
by taking the inverse fast Fourier transform of frequency response H.sub.D
(.omega.) for the digital correlator where H.sub.D (.omega.) is determined
according to the following expression:
##EQU8##
where H.sub.D (.omega.) iS a desired frequency response of said pulse
compressor and H.sub.F (.omega.) is a frequency response of said analog
intermediate frequency filter.
8. The pulse compressor of claim 6, wherein said tap weights are determined
based on a frequency response of said analog intermediate frequency
filter.
9. The pulse compressor of claim 1, wherein said digital correlator
performs digital correlation and pulse compression by weighting phase
components of said digital baseband signals in accordance with phase
weights determined based on the frequency response of said analog
intermediate frequency filter.
10. The pulse compressor of claim 1, wherein said digital correlator has a
frequency response set based on a frequency response of said analog
intermediate frequency filter.
11. The pulse compressor of claim 1, wherein said transmitted pulse is a
non-linear FM waveform.
12. The pulse compressor of claim 1, wherein said transmitted pulse is a
non-linear FM pulse.
13. A method of pulse compression, comprising:
a) receiving echo signals of a transmitted pulse, said transmitted pulse
being an FM signal;
b) filtering and weighting said echo signals with an analog intermediate
frequency filter having a passband less than a frequency sweep of said
transmitted pulse;
c) converting output of said analog intermediate frequency filter into
digital baseband signals; and
d) digitally correlating said digital baseband signals with a digital
correlator to complete pulse compression of said echo signals.
14. The method of claim 13, wherein said step a) performs said filtering
using a Bessel filter as said analog intermediate frequency filter.
15. The method of claim 13, wherein said step d) performs said digital
correlation using a tapped delay line.
16. The method of claim 15, wherein said step d) performs said digital
correlation using tap weights of taps in said tapped delay line which are
determined by taking the inverse fast Fourier transform of frequency
response H.sub.D (.omega.) for the digital correlator where H.sub.D
(.omega.) is determined according to the following expression:
##EQU9##
where H.sub.M (.omega.) is a desired frequency response of said pulse
compressor and H.sub.F (.omega.) is a frequency response of said analog
intermediate frequency filter.
17. The method of claim 15, wherein said step d) performs said digital
correlation using tap weights of taps in said tap delay line which are
determined based on a frequency response of said analog intermediate
frequency filter.
18. The method of claim 13, wherein said step d) comprises the following
steps:
d1) taking a fast Fourier transform of said digital baseband signals;
d2) taking a fast Fourier transform of tap weights of the digital
correlator;
d3) multiplying output of said step d1) and said step d2); and
d4) taking an inverse fast Fourier transform of output from said step d3).
19. The method of claim 18, wherein said step d2) fast Fourier transforms
said tap weights which are determined by taking the inverse fast Fourier
transform of frequency response H.sub.D (.omega.) for the digital
correlator where H.sub.D (.omega.) is determined according to the
following expression:
##EQU10##
where H.sub.D (.omega.) is a desired frequency response of said pulse
compressor and H.sub.F (.omega.) is a frequency response of said analog
intermediate frequency filter.
20. The method of claim 18, wherein said step d2) fast Fourier transforms
said tap weights which are determined based on a frequency response of
said analog intermediate frequency filter.
21. The method of claim 13, wherein said step d) performs digital
correlation and pulse compression by weighting a phase of components of
said digital baseband signals in accordance with phase weights determined
based on a frequency response of said analog intermediate frequency
filter.
22. The method of claim 13, wherein said step d) performs digital
correlation using a digital correlator having a frequency response set
based on a frequency response of said analog intermediate frequency
filter.
23. The method of claim 13, wherein said transmitted pulse is a non-linear
FM waveform.
24. The method of claim 13, wherein said transmitted pulse is a non-linear
FM pulse.
25. A method of using a pulse compressor, comprising;
a) providing an analog intermediate frequency filter with a passband less
than a frequency sweep of a transmitted pulse, said transmitted pulse
being an FM signal;
b) providing a converting means which converts analog signals into digital
baseband signals;
c) providing a digital correlator which digitally correlates baseband data
signals;
d) receiving echo signals of said transmitted pulse with said analog
intermediate frequency filter;
e) filtering and weighting said echo signals with said analog intermediate
frequency filter;
f) converting output of said step e) with said converting means to form
digital baseband data signals; and
g) digitally correlating output of said step (f) using said digital
correlator.
26. The method of claim 25, wherein said step a) provides Bessel filter as
said analog intermediate frequency filter.
27. The method of claim 25, wherein said step c) provides a tapped delay
line as said digital correlator.
28. The method of claim 25, wherein said step c) comprises the steps of:
c1) providing a first FFT circuit for taking a fast Fourier transform of
said digital baseband signals;
c2) providing a second FFT circuit for taking a fast Fourier transformer of
tap weights of said digital correlator
c3) providing a multiplier for multiplying output of said first FFT circuit
and said second FFT circuit; and
c4) providing an inverse FFT circuit for taking an inverse fast Fourier
transform of output from said multiplier.
29. The method of claim 25, wherein said step c) provides said digital
correlator having a frequency response set based on a frequency response
of said analog intermediate frequency filter.
30. A method of manufacturing a pulse compressor, comprising:
a) determining a time function of a transmitted pulse, said transmitted
pulse being an FM signal;
b) determining a desired frequency response of a pulse compressor based on
said time function of said transmitted pulse;
c) selecting an analog intermediate frequency filter with a passband less
than a frequency sweep of said transmitted pulse;
d) determining a frequency response of said selected analog intermediate
frequency filter;
e) providing a converting means which converts analog signals to digital
baseband signals;
f) connecting an input of said converting means to an output of said
selected analog intermediate frequency filter;
g) selecting a digital correlator which digitally correlates and compresses
digital baseband data signals and has a frequency response substantially
equal to said desired frequency response of said pulse compressor divided
by said frequency response of said selected analog intermediate frequency
filter; and
h) connecting an output of said converting means to an input of said
digital correlator.
31. The method of claim 28, wherein said step c) selects an analog
intermediate frequency filter which provides rapid falloff in both
frequency and time domains.
32. The method of claim 28, wherein said step g) selects a digital
correlator which is a tapped delay line.
33. The method of claim 30, wherein said step g) selects a digital
correlator which comprises:
a first FFT circuit taking a fast Fourier transform of said digital
baseband signals;
a second FFT circuit taking a fast Fourier transform of tap weights of said
digital correlator;
a multiplier multiplying output of said first FFT circuit and said second
FFT circuit; and
an inverse FFT circuit taking an inverse fast Fourier transform of output
from said multiplier. |
|
 |
|
 |
Claims |
|
|
|
Description |
|
|
BACKGROUND OF THE INVENTION
This invention relates generally to radar processing systems, and more
particularly to a pulse compression system used therein.
In the art of pulse radar systems, it is well known that the ability of a
radar system to perform detection depends upon the energy content of
transmitted pulses. The larger the energy content, the higher the
signal-to-noise ratio of the returning echoes. A large energy content may
be obtained through pulses with large peak power and/or long pulse
duration. Pulses with small durations or pulse widths are preferred since
the shorter the pulse width the better the range resolution and range
accuracy. The components comprising a radar system, however, often place
limits on the peak power of a pulse and require that pulses of longer
duration be transmitted in order to obtain the necessary energy content
for a pulse. Pulse compression techniques and pulse compressors were
developed to achieve the same resolution of a much shorter pulse with high
peak power using much longer pulses with lower peak power.
The use of pulse compression techniques has become increasingly more
important because pulse compression has made the use of solid state
devices in radar applications feasible. While not being able to produce
short pulses with large peak power, solid state devices can produce lower
powered pulses of long duration.
Initially pulse compressors were implemented as analog devices. These
devices, however, were inflexible with respect to pulse length. To change
the pulse length required changing the device. Additionally, the
components of analog pulse compressors experienced gain drift due to
changes in temperature and aging. Furthermore, the manufacture of some of
these devices was complicated by the fact that each device produced
differed from the other due to deviations during the manufacturing
process. This required additional steps to align each device.
The disadvantages above drove the production of digitally implemented pulse
compressors. Digitally implemented pulse compressors eliminated many of
the disadvantages of the analog pulse compressors such as inflexible pulse
lengths. While initially these systems were very hardware intensive and
expensive to produce, at the present time they are relatively inexpensive
and easy to implement.
An important consideration in the design of digital systems is the sample
rate of the signal being processed. This rate must be high enough to
ensure that the signal is represented accurately without distortion due to
aliasing. At the same time, the higher the rate, the greater the
processing throughput required of the digital hardware, and hence the
greater the cost. Thus it is desirable to minimize the sampling rate under
the constraint of not introducing distortion which degrades the
performance of the system.
Few digital pulse compression system designs constrain or achieve a
constrained sampling rate. U.S. Pat. No. 4,404,562 to Kretschmer describes
one such system. The Kretschmer system transmits a linear FM signal such
as shown in FIG. 1 which has a frequency sweep (bandwidth) of F.sub.2
-F.sub.1 and performs pulse compression on the echoes by means of a
digital pulse compressor 60 illustrated in FIG. 2.
In the Kretschmer device, a signal generator 10 generates an intermediate
frequency signal having a frequency which varies linearly from a frequency
F.sub.1 to a frequency F.sub.2 as time varies from an arbitrary time
t.sub.0 to time t.sub.0 +T. This linearly frequency modulated waveform is
supplied to the mixer 12 via line 14 wherein it is mixed up to radio
frequency (RF) by heterodyning it with an RF signal supplied by an RF
signal generator 16 via line 18. The resultant RF signal is amplified by a
power amplifier 20, passed through a standard duplexer 22, and radiated
from a radar antenna 24.
Echoes received by the antenna 24 are supplied by the duplexer 22 to a
mixer 28 via a line 26. The mixer 28 beats or heterodynes the echo signal
with the RF signals supplied by RF signal generator 16 via line 30 in
order to obtain the intermediate frequency echo signal varying from
F.sub.1 to F.sub.2. The resultant intermediate frequency signal is
amplified in an intermediate frequency amplifier 32 with a bandwidth from
F.sub.1 to F.sub.2 centered on the frequency (F.sub.2 +F.sub.1)/2. Since
the intermediate frequency amplifier 32 passes the same band as the linear
FM waveform, the intermediate frequency amplifier 32 performs noise
reduction but does not perform a pulse compression function.
At this point in the circuit both the transmission and the reception
processing have been analog in nature. The circuit then samples the echo
signal at the Nyquist rate. The Nyquist rate is defined as twice the
reciprocal of the frequency sweep or bandwidth of the linear FM waveform,
or in this case 2/(F.sub.2 -F.sub.1). Where conversion to baseband I and Q
signals is used, Kretschmer contends the Nyquist rate would be 1/(F.sub.2
-F.sub.1) for each baseband I and Q signal. However, an amplifier, such as
intermediate frequency amplifier 32, providing no attenuation between F1
and F2 and high attenuation outside this band is not physically
realizable. Consequently, sampling at 2/(F.sub.2 -F.sub.1) or 1/(F.sub.2
-F.sub.1) for each baseband I and Q signal is not adequate.
To obtain the information baseband, the intermediate frequency echo signal
from intermediate frequency amplifier 32 is beat or heterodyned with a
local oscillator (L.O.) intermediate frequency signal. Accordingly, an I
channel and a Q channel are provided for generating baseband signals and
sampling those signals at the Nyquist rate. The I channel comprises a
multiplier 34 for beating or heterodyning the intermediate frequency echo
signal on line 33 from amplifier 32 with an L.O. intermediate frequency
signal from signal generator 10 via line 35. Likewise, a multiplier 36 in
the Q channel multiplies the intermediate frequency echo signal on line 35
with an L.O. intermediate frequency signal from the signal generator 10
shifted in phase by 90.degree. by the phase shifter 38 and provided via
line 40.
The baseband I and Q signals are then passed through low pass filters 42
and 44, respectively. Kretschmer discusses that these low pass filters may
be optimally adjusted to just pass baseband pulses of length 1/(F.sub.2
-F.sub.1) (see col. 4 lines 30-35 of U.S. Pat. No. 4,404,562). In other
words Kretschmer teaches setting the passbands of low pass filters 42 and
44 equal to the frequency sweep or bandwidth of the linear FM waveform.
Accordingly since the low pass filters 42 and 44 pass the same band as the
linear FM waveform, the low pass filters 42 and 44 perform noise reduction
but do not perform a pulse compression function. The outputs of these low
pass filters 42 and 44 are then applied to sample and hold circuits 46 and
48, respectively.
Kretschmer further teaches that to determine the optimum sampling rate for
the sample and hold circuits 46 and 48 there are two competing factors
which require consideration. In the ideal situation, sampling would begin
at the beginning of the echo pulse. However, no provision in the circuit
can be made for ensuring that sampling will begin at the beginning of the
echo pulse. Accordingly, a sampling error of as much as 1/2 of a sampling
period may exist. Kretschmer teaches reducing the sampling period or
increasing the sampling rate so that the sampling error will be
proportionately reduced (col. 4, lines 37-48 of U.S. Pat. No. 4,404,562).
In the Kretschmer device, however, very high sampling rates reveal the
sidelobe within 13 dB of the mainlobe, which is present in the typical
linear FM response.
The Nyquist sampling rate is the minimum sampling rate which will allow the
Kretschmer circuit to reconstruct all of the information for a given
bandwidth. This rate is generally two times the reciprocal of the
bandwidth, or in the case of baseband I and Q signals it is equal to the
reciprocal of the bandwidth itself. Kretschmer teaches that the use of
Nyquist rate sampling will provide an acceptable sampling error rate and
will also provide a maximized mainlobe to sidelobe ratio (col. 4, lines
57-59 of U.S. Pat. No. 4,404,562). FIG. 5 illustrates the response of the
Kretschmer system with over sampling. As FIG. 5 illustrates, the sidelobe
is only 13 dB down from the mainlobe. When the Nyquist rate is used, as
illustrated in FIG. 3, the sidelobe is 27 dB down from the mainlobe. As
discussed above, the Nyquist rate is 1/(F.sub.2 -F.sub.1) for the baseband
I and Q signals. Accordingly, the sample and hold circuits 46 and 48 are
driven at this rate. The sampling pulses are supplied via line 50 from the
signal generator 10. The sample and hold circuits 46 and 48 hold their
sample values for a time 1/(F.sub.2 -F.sub.1) between the samples.
The I and Q samples from the sample and hold circuits 46 and 48 are used to
modulate multipliers 52 and 54, respectively. An L.O. intermediate
frequency signal is supplied to multiplier 52 in the I channel which will
operate to modulate that intermediate frequency signal with the sampled I
signal from the sample and hold circuit 46. Kretschmer teaches that this
intermediate frequency signal may be the signal F.sub.1 supplied via line
35. Likewise, the multiplier 54 in the Q channel is supplied with an L.O.
intermediate frequency signal in quadrature with the intermediate
frequency signal supplied to the multiplier 52. This quadrature L.O.
intermediate frequency signal is modulated by the sampled Q output signal
from the sample and hold circuit 48. Again, Kretschmer teaches this L.O.
intermediate frequency signal may be the L.O. intermediate frequency
signal F.sub.1 shifted in phase by 90.degree. and supplied via line 40.
The intermediate frequency signal output from the multipliers 52 and 54
are then added together by an addition circuit 56 and the sum supplied to
a compression circuit 60. The purpose of the compression circuit 60 is to
take successive samples in time, and weight those samples such that when a
received signal is properly indexed in the circuit, its output will be a
short pulse with a significant amplitude, i.e., a compression operation.
The compression circuit 60 is a tap delay line, whose length will, of
course, be determined by the uncompressed length T of the transmitted
pulse. The number of taps on the delay line is generally determined by the
number of samples taken in the sample and hold circuits 46 and 48. The
delay line is composed of a series of p-1 cascaded delay elements 62
each equal to a delay of .tau.=T/p wherein T/p =1(F.sub.2 -F.sub.1). A
signal tap 64 is taken before each delay element 62 and a final tap 65 is
taken after the last delay element for a total of p signal taps. The delay
elements 62 may be formed by cable or standard RC transmission line cut to
the proper length. Equal amounts of signal will be obtained from each tap
by setting the tap impedances in the well known manner.
If the originally transmitted signal had contained a single frequency
across the length of the pulse, then the signals from these p taps could
be added without further processing. However, because the frequency varies
with time during the length of the pulse, the signals on the individual
signal taps must be progressively phase shifted back into phase.
Accordingly, in order to bring the signals on the various signal taps into
phase with each other, phase weighting elements 66, 68, 70 and 72 are
provided for the different signal taps. The phase weights to be set in
these individual phase weighting elements is determined in the well known
manner as follows. For purposes of the present discussion, the phase of
the signal on the last signal tap 65 will be taken as the reference.
Accordingly, the phase weighting element 66 will provide a phase shift of
zero. The phase shifts for the other phase weighting elements may be
calculated as follows:
The frequency difference between the signals on any two taps f.sub.diff is
##EQU1##
For a linear frequency modulation, as in this instance, df/dt=a constant k
or
f.sub.diff =kt (2)
also
##EQU2##
.PHI. diff taps
##EQU3##
The constant k is found to equal the slope of the frequency vs time
transmission characteristic or (F.sub.2 -F.sub.1)/T=B/T. Thus, the phase
weighting elements will provide the following phase shifts
##EQU4##
These phase weighting elements may comprise BNC cable or twisted pairs cut
to the proper length in order to obtain the prescribed phase shifts.
When the linear variation of the frequency of the signal with time has been
taken into account, then the output from the phase weighting elements 66
through 72 should all be in phase when a received echo pulse is properly
indexed in the delay line. Accordingly, the weighted signal output from
the phase weighting elements 66 through 72 are added together in an adding
circuit 74.
As the discussion above reveals, the Kretschmer device may only be realized
in the ideal world of computer simulations because the filter
characteristics specified by Kretschmer are not physically realizable.
Furthermore, Kretschmer discusses that a sampling error of as much as 1/2
the sampling period may exist. FIG. 3 illustrates the response of the
Kretschmer device when there is no sampling error, and FIG. 4 illustrates
the response of the Kretschmer device when a sampling error of 1/2 a
sampling period exists. A comparison of FIGS. 3 and 4 demonstrates that
the peak response of the Kretschmer device degrades by 3.9 dB and the
width of the main lobe widens significantly when such a sampling error
exists. Therefore, as the sampling error increases, the Kretschmer device
will be unable to resolve the closely spaced targets that were resolved
when no sampling error was present. In the radar art, loss due to sampling
error is called range sampling loss, and is often expressed as the
difference between the peak signal-to-noise ratio and the signal-to-noise
ratio averaged over all possible sampling points. Kretschmer's approach
results in a range sampling loss of 1.5 dB.
As further discussed above, Kretschmer teaches that the performance of his
device depends on the sampling rate. Specifically, reducing the sampling
rate proportionally reduces the sampling error discussed above. FIG. 5
illustrates the deleterious affects of over sampling (sampling above the
Nyquist rate) with the Kretschmer device. As illustrated in FIG. 5, the
sidelobe is only 13 dB down from the mainlobe. When the Nyquist rate is
used, as illustrated in FIG. 3, the sidelobe is 27 dB down from the
mainlobe. The closer the sidelobe becomes to the mainlobe, the greater the
chance of erroneous or missed target detection. For example, the sidelobes
produced by clutter such as mountains or other targets could mask the
mainlobe of a target and prevent detection of that target. Accordingly,
Kretschmer as discussed previously mandates the use of the Nyquist
sampling rate, the lowest possible sampling rate, as the sampling rate of
the sample and hold circuits 46 and 48 so that the sidelobe weighting
shown in FIG. 3 is obtained.
The Kretschmer device suffers from other disadvantages with respect to
sidelobe suppression and the flexibility of the transmitted waveform. FIG.
6 illustrates the response of the Kretschmer device with sidelobe
weighting, and is a copy of FIG. 7 in U.S. Pat. No. 4,404,562 discussed in
col. 7 lines 36-42. Based on FIG. 6, the peak response of the Kretschmer
device is about -3 dB, which represents a mismatch loss of 3 dB. As
discussed previously, resolution and the range at which targets are
detectable depend upon the energy content of the transmitted pulse.
Consequently, the greater the mismatch loss, the greater the energy
content of the transmitted pulse must be to overcome such a loss and
obtain the desired range and resolution. A mismatch loss of 3 dB would
require doubling the output power of the transmitter. In the real world,
such power requirements are not economically feasible; especially with
solid state devices. As such, a mismatch loss of 3 dB represents an
unacceptably high mismatch loss for a practical system.
Additionally, the Kretschmer device can use only a linear FM waveform such
as shown in FIG. 2 as the transmitted pulse. Consequently, the Kretschmer
pulse compressor is not applicable or flexible enough to be used in a
radar system using a non-linear FM waveform as the transmitted pulse.
Furthermore, Kretschmer claims that the device of U.S. Pat. No. 4,404,562
is Doppler tolerant. FIG. 7 illustrates the response of the Kretschmer
device with Doppler. The Doppler chosen represents a high-speed target at
L-band. A comparison of FIG. 7 to FIG. 3 shows a peak response reduction
of 0.5 dB, and a significant broadening of the mainlobe. This,
particularly the broadening of the mainlobe, adversely impacts the
resolution of the Kretschmer system.
SUMMARY OF THE INVENTION
An objective of the present invention is to provide an apparatus and method
of pulse compression which overcomes the above noted disadvantages.
Another objective of the present invention is to provide an apparatus and
method of pulse compression which minimizes mismatch loss while maximizing
sidelobe suppression.
A further objective of the present invention is to provide an apparatus and
method of pulse compression having a response substantially unaffected by
sampling error.
An additional objective of the present invention is to reduce the sampling
rate to very close to the Nyquist rate, which reduces the cost of the A/D
convertor, the digital pulse compressor, and primarily all subsequent
digital signal processing, such as, for example, Doppler filtering.
Another objective of the present invention is to provide an apparatus and
method of pulse compression having increased flexibility with respect to
the transmitted pulse.
A further objective of the present invention is to provide an apparatus and
method of pulse compression which is Doppler tolerant.
A further objective of the present invention is to provide a method of
using a pulse compressor which overcomes the above noted disadvantages and
achieves the above discussed objectives.
An additional objective of the present invention is to provide a method of
manufacturing a pulse compressor which overcomes the above noted
disadvantages and achieves the above discussed objectives.
The above and other objectives are achieved by providing a hybrid
analog-digital pulse compressor, comprising an analog intermediate
frequency filter receiving echo signals of a transmitted pulse and having
a passband less than a frequency sweep of said transmitted pulse, said
analog intermediate frequency filter filtering and weighting said echo
signals; converting means for converting output of said analog
intermediate frequency filter into digital baseband signals; and a digital
correlator for digitally correlating said digital baseband signals to
complete pulse compression of said echo signals.
These objectives are further achieved by providing a method of pulse
compression, comprising a) receiving echo signals of a transmitted pulse;
b) filtering and weighting said echo signals with an analog intermediate
frequency filter having a passband less than a frequency sweep of said
transmitted pulse; c) converting output of said analog intermediate
frequency filter into digital baseband signals; and d) digitally
correlating said digital baseband signals with a digital correlator to
complete pulse compression of said echo signals.
These objectives are also achieved by providing a method of using a pulse
compressor, comprising a) providing an analog intermediate frequency
filter with a passband less than a frequency sweep of a transmitted pulse;
b) providing a converting means which converts analog signals into digital
baseband signals; c) providing a digital correlator which digitally
correlates baseband data signals; d) receiving echo signals of said
transmitted pulse with said analog intermediate frequency filter; e)
filtering and weighting said echo signals with said analog intermediate
frequency filter; f) converting output of said step e) with said
converting means to form digital baseband data signals; and g) digitally
correlating output of said step (f) using said digital correlator.
These objectives are further achieved by providing a method of
manufacturing a pulse compressor, comprising a) determining a time domain
structure of a transmitted pulse; b) determining a desired frequency
response of a pulse compressor based on said time domain structure of said
transmitted pulse; c) selecting an analog intermediate frequency filter
with a passband less than a frequency sweep of said transmitted pulse; d)
determining a frequency response of said selected analog intermediate
frequency filter; e) providing a converting means which converts analog
signals to digital baseband signals; f) connecting an input of said
converting means to an output of said selected analog intermediate
frequency filter; g) selecting a digital correlator which digitally
correlates and compresses digital baseband data signals and has a
frequency response substantially equal to said desired frequency response
of said pulse compressor divided by said frequency response of said
selected analog intermediate frequency filter; and h) connecting an output
of said converting means to an input of said digital correlator.
Other objects, features, and characteristics of the present invention;
methods, operation, and functions of the related elements of the
structure; combination of parts; and economies of manufacture will become
apparent from the following detailed description of the preferred
embodiments and accompanying drawings, all of which form a part of this
specification, wherein like reference numerals designate corresponding
parts in the various figures.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 illustrates the frequency sweep of a linear FM signal transmitted
according to the prior art;
FIG. 2 illustrates a prior art pulse compressor;
FIG. 3 illustrates the response of the prior art pulse compressor of FIG. 2
without sampling error;
FIG. 4 illustrates the response of the prior art pulse compressor of FIG. 2
with sampling error;
FIG. 5 illustrates the deleterious effects of over sampling on the prior
art pulse compressor of FIG. 2;
FIG. 6 illustrates the response of the prior art pulse compressor of FIG. 2
with sidelobe weighting;
FIG. 7 illustrates the response of the prior art pulse compressor of FIG. 2
with Doppler;
FIG. 8 illustrates a pulse compressor according to a first embodiment of
the present invention;
FIG. 9 illustrates a pulse compressor according to a second embodiment of
the present invention;
FIG. 10 illustrates an alternative embodiment for the digital correlator in
the first and second embodiments of the present invention;
FIG. 11 illustrates the frequency sweep of a nonlinear FM waveform
transmitted according to the present invention;
FIG. 12 illustrates the response of the pulse compressor of the present
invention without sampling error;
FIG. 13 illustrates the response of the pulse compressor of the present
invention with sampling error;
FIG. 14 illustrates the affects of over sampling on the pulse compressor of
the present invention;
FIG. 15 illustrates the response of the pulse compressor of the present
invention with Doppler;
FIG. 16 illustrates the impulse response of a Bessel intermediate frequency
filter according to the present invention;
FIGS. 17 and 18 illustrate the amplitude and phase values for the tap
weights of the digital correlator according to the present invention;
FIG. 19 illustrates the oversampled theoretical response of the pulse
compressor according to the present invention; and
FIG. 20 illustrates the measured response of an actual radar system which
employs the pulse compressor according to the present invention.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
FIG. 8 illustrates a radar system with the hybrid analog-digital pulse
compressor of the present invention. In FIG. 8, a waveform synthesizer 100
generates a non-linear FM waveform. While the present invention will be
described with respect to the generation of a non-linear FM waveform,
waveform synthesizer 100 can also, alternatively, generate a linear FM
waveform. The instantaneous frequency of the non-linear FM waveform
generated by the waveform synthesizer 100 is illustrated in FIG. 11. The
non-linear FM waveform has a pulse length L and a frequency sweep
(bandwidth) of F.
Two up-converter mixers 102 are connected in series to the waveform
synthesizer 100. Each up-converter mixer 102 mixes or heterodynes the
non-linear FM waveform with an RF signal to convert the non-linear FM
waveform to radio frequency. In this embodiment, double conversion (i.e.,
two up-converters 102) are used.
The coherent oscillator 104 and stable local oscillator 108 are connected
to respective up-converter mixers 102. Each of coherent oscillator 104 and
stable local oscillator 108 generates an RF signal which is mixed with the
non-linear FM waveform by a respective one of the up-converter mixers 102
to up-convert the non-linear FM waveform to radio frequency.
The output of the final up-converter mixer 102 is supplied to power
amplifier 20, which amplifies the non-linear FM waveform. A standard
duplexer 22 receives the output of power amplifier 20, and passes the
output of power amplifier 20 to an antenna 24. The antenna 24 then
radiates the up-converted and amplified non-linear FM | | |
