GB2279196A - Processing means for radar system - Google Patents

Abstract

In a processing mean for a S.A.R. radar system, in which radar returns corresponding to a stationary target are integrated after phase or frequency slope compensation 12 corresponding to the tangential velocity of the aircraft and the range, such stationary targets are distinguished. Additional phase compensation channels 12a to 12e are provided applying different frequency slopes (k2 - k2 Fig 12) leading to phase corrections corresponding to different Doppler frequency variation rates. The integration peaks that correspond to targets are now compared for the five channels (Fig 13), and the comparison enables moving targets, coherent stationary targets and, incoherent distributed stationary targets (i.e. clutter) to be distinguished from each other. The channels 12 and 12a to 12e are provided for each range gate, 10, 11. <IMAGE>

Description

PROCESSING MEANS FOR RADAR SYSTEM This invention relates to processing means for radar systems.

The invention is especially applicable to processing means including means for integrating the signal energy from groups of radar returns, such as in S.A.R. (synthetic aperture radar).

A known processing means for S.A.R. is shown in figure 1. The transmitter 1 feeds pulses 2 of a waveform 3 (figure 3) to an antenna 4 via a duplexer 5. The S.A.R. is mounted on an aircraft 6 (figure 2) travelling at a velocity v and looks sideways relative to its flight track AB.

Between pulses, echo returns 7 received from targets are fed via the antenna 4 and duplexer 5 to a receiver 8, which receives as input a local oscillator 9 coupled by a loop coupler 9a to the transmitter, so that the phase of the radar returns as well as the amplitude can be determined.

In practice, the aircraft tracks will depart from AB by varying distances over a period of time, and the aircraft will rotate through varying angles such that AB is its mean fore and aft direction. Such rotational and translation motions are compensated for.

Typically, a target such as T will be illuminated by the beam (figure 2) for around 40 seconds as the beam 9b sweeps over the target. The amplitude response of the antenna is shown in figure 4: the radar returns received when the target is near the edge of the beam width have a much lower amplitude than those received when the target is on boresight.

To increase the signal to noise ratio, returns 7 for each particular range cell of interest are integrated over a period of time e.g. 2 seconds and many thousands of pulses are received over this time. However, the radar returns cannot be simply added together because the phase of a radar return from a stationary single point target relative to that of the transmitted pulse varies across the width of the beam (figure 5) in a somewhat parabolic form. Corresponding to this, the frequency of the radar returns relative to the outgoing pulses is of approximately linear form (figure 6), frequency being the differential of phase with respect to time.The change in single frequency from +fas to -fs as the antenna illumination beam moves over the target is due to the Doppler effect as the aircraft first of all approaches and then recedes from the target. The signal frequencies and and -fs corresponding to the edges of the beam and also the rate at which the frequency changes from +fas to -fs both depend on the velocity v of the aircraft.

In the known processing means of figure 1, the phase variation is compensated for by producing a compensating frequency variation (sweep 13 - figure 6) in the receiver in circuit 12, the frequency versus time slope of which is the same in magnitude but opposite in sign to that of the input signal. The sweep is centred on zero frequency at the time when the target represented by the input signal fs to is on boresight (i.e. time to) and the duration is the integration interval.

When the compensating frequency 13 is mixed with the input signal radar returns, the frequency slope of the input signal is corrected to zero frequency, and the returns are now all at a constant phase and give maximum amplitude when integrated and detected. If the compensating frequency was centred on zero when the target was marginally off boresight (13'), the returns would be modulated in amplitude by the beat frequency produced on mixing with the input signal frequency. It follows that the sum of integrated returns would now be less than the sum corresponding to the time of the target beam on boresight.

It is not of course known when the target is on boresight, and so integrations are carried out very frequently (typically at 0.05 second intervals - figure 7) and the output of the integrator 14 passes through a peak at the time that the targets T1, T2 are on boresight (figure 8). Each line 13 represents frequency compensation and integration over a two second period. The detector 15 produces a measure of the integration amplitude and is followed by an amplitude comparison which can be set to give an output if it exceeds a threshold level that is set to clear the peaks of noise fluctuation.

The input signals from all stationary targets will have the same frequency slope as the beam sweeps across the target, the slope depending on the velocity of the aircraft relative to the target and on the range of the target. If the target was fast-moving towards the aircraft, it would have the same frequency slope as for a stationary target, but would be shifted relative to the frequency slope for a stationary target by a Doppler frequency fd appropriate for the velocity of approach on the target (line 16 - figure 9).

In the Applicants prior UK patent application no.

8614824, additional receiving channels similar to that shown in figure 1 were provided, except that each was prefaced by an SSB (single side band) mixer tuned to a different frequency shift, in order to shift the Doppler shifted frequency slope 16 back to pass through zero frequency at some time within the antenna illumination period of the moving target, and therefore to enable signal integration to be performed to enable a target to be detected. From a single such detection it would be possible to distinguish a fast-moving target in the radial direction from stationary targets. However, it is necessary to compare the detected amplitudes from at least two of the receiving channels in order to determine the Doppler frequency of the target, and to determine the time (to) at which it appeared in the centre of the antenna beam.Targets moving slowly in the radial direction are difficult to detect by this method since their range of frequency variation overlaps that of stationary targets.

The input signals of a target with a tangential velocity utangental or a radial acceleration bradial would have a different slope 17 to that of a stationary target (figure 9).

Referring to figure 10, the input signals of a succession of targets at the same range are illustrated, the target T3 being on boresight at time t03 etc. Targets T3, T5 are stationary, target T8 is fast moving towards the aircraft and target T9 slow moving away from the aircraft: all have the same slope. Targets T4 T6 have a finite tangential velocity or radial acceleration and have different slopes.

While the techniques of the Applicants prior UK patent application can be relied upon to distinguish fast radially-moving targets T8 from stationary targets, since a fast radially-moving target is outside the frequency sweep corresponding to the stationary targets, it is unable to distinguish slow radially-moving targets such as T9 or tangentially moving or radially accelerating targets such as T4, T6 from stationary targets since their frequency sweeps lie within, at least partly, the frequency sweep of stationary targets. Further, such moving targets (eg a truck moving along a road) may be submerged by the stationary target spectrum if large amplitude stationary targets are present. Further, the problem is aggravated if the radar antenna beam width is large, which causes the stationary target spectrum to increase and occupy a large part of the overall moving target spectrum.

The invention provides processing means for a radar system, comprising means for applying more than one phase compensation characteristic to groups of radar return pulses corresponding to different Doppler frequency variation rates and means for integrating signal energy from differently phase compensated groups of radar returns.

Integrating signal energy from differently phase compensated groups of radar returns enables moving targets to be distinguished from stationary targets.

Processing means for a S.A.R. constructed in accordance with the invention will now be described by way of example with reference to the accompanying drawings, in which: Figure 11 is a schematic block diagram of processing means; Figure 12 illustrates schematically differently compensating frequencies applied to the input signal; Figure 13 illustrates schematically the detected outputs of integrators corresponding to the different frequency compensations.

Referring to figure 11, a transmitter 1 mounted on an aircraft as in figure 2 emits pulses of the form shown in figure 3 via a duplexer 5. The radar returns, as in the prior processing means of figure 1, pass to a receiver 8, in which amplitude and phase shift across the beam width are determined by means of a local oscillator 9 which is made to keep in step with the transmitted waveform frequency which is picked up by a loop coupler 9a.

As in the known processing means, the processing means of the invention has a plurality of channels for providing frequency slope compensation to the stationary targets, there being one channel for each range gate 10, and only one channel being illustrated. Each channel comprises: a circuit 12 providing frequency sweeps of the same magnitude but of opposite sign of that of stationary targets, centred about zero frequency and of the duration of the integration interval (e.g. two seconds): an integrator 14 for carrying out integrations; and a detector 15 for detecting peaks in the integrations corresponding to targets as shown in figure 7 and 8. The frequency sweeps, the integrations and the detections are carried out repeatedly (e.g. at 0.05 second intervals).

In accordance with the invention, the processing means is provided with a plurality of sets of additional channels, one set for each range gate 11, and only one set being illustrated. Each set has a number of channels e.g. five, 12 a-e, 14 a-e and 15a, which differ from channel 12, 14, 15 only in the slope applied to the input signal to provide phase compensation. Thus, while the circuit 12 provides frequency compensation ko to provide zero beat frequency of the Doppler frequency variation rate of a stationary target, the channels 12a-e provide frequency compensation k0 and two higher and two lower values of frequency compensation k1, k2, k1, k.2 k2 (figures 11 and 12).

There are a greater number of range gates 10 for the stationary target detection channel, to produce high resolution, than range gates 11 for the sets of additional channels, where coarse resolution may be used since there will be relatively few moving targets.

Thus, if an actual target is stationary, the result of mixing the input signal with compensation frequency ko in circuit 12 will produce a continuous zero beat frequency and the radar pulse returns are all at a constant phase and give maximum amplitude when integrated and detected. The result of mixing the input signal frequency with the compensation frequencies of circuits 12a-e will produce a time-varying beat frequency and the amplitudes of the radar returns will have that time-varying beat frequency superimposed on them, so that they will cancel to some extent when integrated, producing respective lower outputs. Each integration referred to with reference to figure 7 is now carried out five times.Although there are various ways of assessing the results, one way is to set the detected amplitude comparison threshold such that each channel only produces an output when a threshold corresponding to an integrator output is exceeded, and then these outputs (each corresponding to the peak of the profile of figure 8) are compared to ascertain which is the largest. In the case of the stationary target, the output from channel 12c is greatest. In the case of a target moving with either a component of radial acceleration or of tangential velocity, or both, the output from one of the other channels 12a, b, d, e will be the largest. This will occur irrespective of whether the target has a zero or finite component of radial velocity.

A better estimate of the target acceleration or tangential velocity can be obtained by plotting the peak outputs of the five channels corresponding to a target on curves 16a-e representing the outputs of the integrators at different accelerations or tangential velocities. If the peak output from integrator 14c corresponding to circuit 12c is at point A and the peak output from integrator 14b corresponding to circuit 12b is B, then the acceleration or tangential velocity can be interpolated as C. In this way, a moving target can be distinguished from a stationary target.

Background clutter, on the other hand, will be phase incoherent, and will therefore give similar outputs in all slope channels, and will therefore be distinguished by this similarity.

Naturally the same procedure applies to potential targets in each of the range cells of interest. In order to economise on processing complexity i.e. the number of signal integrations in a given time, it is permissible to join together (i.e. to sum or integrate) several range cells of the moving target channels before applying frequency slope compensation. The signal/noise ratio will be reduced by a small amount by so doing, so this economy will not normally be carried in the stationary target channel, where optimum range and azimuth resolution are usually required.

If it is desired to measure the target motion more accurately, or to determine the instant that the target appeared in the centre of the antenna beam (two) and thus to locate the moving target in the azimuth direction on a radar map, the following procedure may be adopted. Following a moving target detection, the signal velocity or acceleration and range being approximately known, the signal is then directed to several finer velocity and acceleration measurement channels to determine the velocity and acceleration more accurately. This is therefore a cueing method, whereby initial detections cue in more sophisticated processing and is possible since there will not be many moving target detections over a short period. The same procedure is carried out in at least one other set of five channels for a fixed frequency shift, in the manner previously described for fast radially-moving targets, i.e.

after each range gate 11, one or two further groups of the five channel set 12a-e etc are provided, each prefaced by an SSB mixer tuned to a different fixed frequency shift, in the manner described previously for fast moving targets in the prior UK patent application No. 8614824. By comparing the peak amplitudes in the two or more five-channel sets, the Doppler frequency fd corresponding to the radial velocity of the target may be determined, and the time at which the target was in the centre of the antenna beam may also be determined from the different times of detection in the SSB frequency shifted set of channels. By using the extra channels of different Doppler variation rates in the manner described, the detection and location of targets with a low value of radial velocity is now possible.

The ability to distinguish background clutter from other types of target is valuable, enabling the removal of distributed background clutter effects i.e. speckle from a radar map. Alternatively background clutter may be displayed as such if required. Moving targets may be displayed in the correct locations on the same map. This is a useful facility and would enable the radar to indicate when a vehicle is moving and along which road.

The antenna shaping (figure 4) could be used to resolve the ambiguity between radial acceleration and tangential velocity. The time-width of the antenna shaping is affected by tangential velocity but not by radial acceleration (or radial velocity either).

Means will be provided for compensating for translational and rotational motions of the aircraft antenna. A suitable value for the pulse repetition frequency is 10 kHz, and a suitable pulse duration is 13 nanoseconds.

Variations may be made without departing from the scope of the invention. Thus, while the frequency slope compensation has been described for a circuit which may be executed in an analogue fashion, the signal from the receiver may be digitised and the frequency slope compensation may be carried out as a phase compensation, and may be carried out in the integrator. Although the processing has been described in relation to S.A.R., it may also be used in other situations where signal energy from a number of returns is integrated, such as in Doppler beam sharpening.

Claims (5)

1. Processing means for a radar system, comprising means for applying more than one phase compensation characteristic to groups of radar return pulses corresponding to different Doppler frequency variation rates, and means for integrating the signal energy from differently phase compensated groups of radar returns.

2. Processing means as claimed in claim 1, in which frequency slope compensation is carried out by mixing the input signal with a frequency sweep.

3. Processing means as claimed in claim 1 or claim 2, in which the integrating means is arranged to perform several simultaneous integrations of the input signal energy using phase compensations at respective different Doppler frequency variation rates, and the peaks of the integrations for the different Doppler frequency variation rates are compared with each other.

4. Processing means as claimed in any one of claims 1 to 3, in which the radar system is a S.A.R.

5. Processing means for a radar system substantially as herein described with reference to Figures 11 to 13 of the accompanying drawings.

5. Processing means for a radar system substantially as herein described.

Amendments to the claims have been filed as follows

1. Processing means for a radar system, comprising means for applying more than one phase compensation characteristic corresponding to more than one Doppler frequency variation rate to groups of radar return pulses to produce more than one phase compensation for each group of returns, and means for integrating the signal energy from each group of radar returns for each of the different phase compensations, to enable moving targets to be distinguished from stationary targets.

2. Processing means as claimed in claim 1, in which frequency slope compensation is carried out by mixing the input signal with a frequency sweep.

3. Processing means as claimed in claim 1 or Claim 2 in which the integrating means is arranged to ?-r-for- several simultaneous integrations of the input siallal ers-ç using phase compensations at respective different doppler frequency variation rates, and the peaks of the integrations for the different Doppler frequency variation rates are compared with each other.

4. Processing means as claimed in any one of claims 1 to 3, in which the radar system is a S.A.R.