GB2248359A - Resolving range/Doppler coupling in a swept-frequency radar - Google Patents

Abstract

Swept-frequency radar apparatus comprises a swept- frequency oscillator (1) the output (2) of which is coupled to an aerial (3) and to one input (7) of a mixer (6). The other input (5) of the mixer is fed from the aerial, so that the mixer generates a boat signal the frequency of which has components due to the range and to the relative velocity (if any) respectively of a reflecting target - so-called range- Doppler coupling. This frequency is analysed by means of a Discrete Fourier Transform (DFT) calculating circuit (23). In order to resolve the range-Doppler coupling the frequency sweeps produced by the oscillator take on in succession repetition rates R1, R2, ... the ratios between which are reducible to the quotient of relatively prime integers, Ni sweeps being generated at each rate Ri, where the ratios between the Ni are equal to or reducible to the corresponding quotients (Fig 2B). The Ni successive complex amplitudes occurring at a given spectral component output (29) of the DFT calculating circuit (23) for each respective repetition rate Ri, are discrete Fourier transformed by a respective circuit (43, 45, 47) to yield from each a respective possibly ambiguous representation of the component due to the target velocity. The ambiguity is then resolved by taking these respective frequencies together (55) so that the true range and/or velocity can be calculated. The lower rate sweeps may be terminated at a frequency corresponding to that reached at the highest rate. <IMAGE>

Description

DESCRIPTION RESOLVING RANGE/DOPPLER COUPLING IN A SWEPT-FREQUENCY RADAR This invention relates to a method of resolving range/doppler coupling in a swept-frequency radar apparatus, in which method the apparatus transmits a succession of segments of radio frequency signal during each of which the frequency of the signal sweeps from a first given value to a second given value at a rate which is the same for each sweep, the transmitted segments, after reflection by said target back to the apparatus, are each mixed with a sample of the signal presently transmitted to yield a beat signal the frequency of which has components due to the range and to the relative velocity (if any) respectively of the target, and the component due to the relative velocity is determined from the beat signal obtained from a plurality of the reflected segments.The invention also relates to radar apparatus for performing such a method.

Radar apparatuses which transmit a succession of segments of radio frequency signal during each of which the frequency of the signal sweeps from a first given value to a second given value at a rate which is the same for each sweep, and mix each transmitted segment, after reflection by a target, with a sample of the signal presently transmitted to yield a beat signal are well-known. If the relative velocity of the radar apparatus and target is zero the frequency of the beat signal is representative of the range of the target, because the beat frequency is equal to the change in the frequency of the transmitted signal which occurs during the time taken for radio signals to travel to the target and back again.

the rate of change of frequency of the transmitted signal is known, said time, and hence the range of the target, can be calculated.

However, if the relative velocity of the radaP-apparatus and target is non-zero the instantaneous frequency of each reflected segment will be Doppler-shifted, resulting in the beat frequency being itself shifted; the beat frequency is now representative of a combination of the range and the relative velocity of the target so-called range - Doppler coupling.

One known way to resolve the range - Doppler coupling is to also transmit segments of radio-frequency signal during each of which the frequency of the signal sweeps in the opposite direction, i.e. from the second given value to the first given value, at the same rate, because for these segments the Doppler effect on the resulting beat frequency will be in the opposite direction. If this is done a given target will therefore give rise to two beat frequencies in all, the mean of which is representative of the target range and the separation between which is representative of the target velocity.However, as discussed in the paragraph at lines 28-48 of page 1 of GB-A-2172461 (PHB33150), it can be difficult to ascertain which of a multitude of beat frequencies which will often be present in practice constitute a related pair, i.e. relate to the same target, for this purpose.

As is also mentioned in GB-A-2172461 at lines 23-27 of page 1, an alternative way of resolving the range - Doppler coupling is, in principle, to use frequency sweeps in one direction only and observe the beat frequency arising from a given target over a sufficient period of time for any change in this beat frequency due to a change in target range to become apparent. The relevant passage includes a comment to the effect that such an approach is unpracticable because the "sufficient period of time" can be considerable, particularly with relatively slow-moving targets. It is an object of the present invention to mitigate this disadvantage.

According to one aspect the invention provides a method as defined in the first paragraph which is characterized in that the repetition rate of said segments takes on in succession first and second values R1 and R2 respectively the ratio between which is reducible to the quotient of first and second integers which are relatively prime, the numbers of segments transmitted at the repetition rates R1 and R2 are at least N1 and N2 respectively where the ratio N1::N2 is equal to or reducible to said quotient, for each said repetition rate Ri an Ni-point Discrete Fourier Transform is taken of the complex amplitudes of the beat signals resulting from Ni successive reflected segments to yield a respective frequency which is a possibly ambiguous representation of said component due to the relative velocity, and the possible ambiguity is reduced or resolved by taking said respective frequencies together.

It has now been recognised that the beat signal resulting from each reflected segment contains a phase component which, if the target has a non-zero velocity relative to the radar apparatus, advances or retards from segment to segment at a rate which is representative of this velocity. In fact, if the segments have a repetition rate R, the phases of the beat signals resulting from successive reflected segments advance or retard at a rate which is equivalent to sampling the Doppler frequency component of the reflected segments at a rate R.Thus, if these phases are determined for N successive reflected segments, e.g. by taking a respective Discrete Fourier Transform of the beat signal resulting from each successive segment, and then a further Discrete Fourier Transform is taken of the N complex amplitudes thus obtained for a given frequency component of the beat signals (which signals may contain more than one frequency component due to the presence of more than one target), the result will be a frequency which is in principle the Doppler frequency component of the relevant beat frequency component, allowing the range - Doppler coupling to be resolved for the relevant target.However, said result is only "in principle" said Doppler frequency component because it is possible that the Doppler frequency component is of a higher frequency than the segment repetition rate R, in which case it will be aliassed down to within the frequency range zero to R, so that for a single repetition rate R the said result is ambiguous. Such ambiguities can, however, be reduced or resolved by employing in succession first and second segment repetition rates R1 and R2 respectively the ratio between which is reducible to the quotient of first and second integers which are relatively prime, the final result then being unambiguous provided that the Doppler component frequency is not greater than the product of these integers and any factor which is common to all the repetition rates employed.The Chinese Remainder Theorem or another decoding arrangement having a similar function can be used to calculate the said final result from the respective results obtained for the various repetition rates (more than two rates may be employed in practice) provided that said results are each expressed as an integer number of sub-multiples of the relevant repetition rate Ri, and that these sub-multiples are the same size for each repetition rate R employed.When, for example, just two repetition rates R1 and R2 are employed, these provisos can be satisfied by arranging that the numbers of segments transmitted at the repetition rates Ri and R2 are at least N1 and N2 respectively, where the ratio N1:N2 is equal to or reducible to said quotient, and that for each said repetition rate Ri the aforesaid further Discrete Fourier Transform is an Ni-point transform taken of the complex amplitudes of the beat signals resulting from Ni successive reflected segments.Because, as is known, the resolution of a Discrete Fourier Transform is directly proportional to the length of the transform (number of points) and the length of each transform is made proportional to the corresponding segment repetition rate, the frequency resolution of each transform is the same, as required. (The Doppler frequency bins coincide for the two repetition rates employed). The said provisos can be satisfied in a similar manner when more than two sweep repetition rates are employed.

According to another aspect the invention provides radar apparatus comprising a swept-frequency RF oscillator the output of which is coupled to an aerial, which oscillator is constructed to generate for transmission a succession of segments of radio-frequency signal during each of which the frequency of the signal sweeps from a first given value to a second given value at a rate which is the same for each sweep, a mixer having inputs coupled to the output of the oscillator and to an aerial respectively, for mixing each transmitted segment, after reflection by a target, with a sample of the signal presently transmitted to yield a beat signal the frequency of which has components due to the range and to the relative velocity (if any) respectively of the target, and means for determining the component due to the velocity (if any) from the beat signal obtained from a plurality of the reflected segments, characterized in that the oscillator is constructed to generate for transmission segments the repetition rate of which takes on in succession first and second values R1 and R2 respectively the ratio between which is reducible to the quotient of first and second integers which are relatively prime and in such manner that the numbers of segments transmitted at the repetition rates R1 and R2 are at least N1 and N2 respectively where the ratio N1::N2 is equal to or reducible to said quotient, and in that said means comprises means for taking, for each said repetition rate Ri, an Ni-point Discrete Fourier Transform of the complex amplitudes of the beat signals resulting from Ni successive reflected segments to yield a respective frequency which is a possibly ambiguous representation of said component due to the relative velocity, and for reducing or resolving the ambiguity by taking said respective frequencies together.

An embodiment of the invention will now be described, by way of example, with reference to the accompanying diagrammatic drawings in which: Figure 1 is a diagram of the embodiment, and Figure 2 at a,b,c,d,e,f,g,h,iand k illustrates various signals which occur in the embodiment of Figure 1.

In Figure 1 a radar apparatus of the continuous-wave swept-frequency type comprises in conventional manner a voltage-controlled variable frequency RF oscillator 1 the output 2 of which is coupled to a transmit/receive aerial 3 via a circulator 4. Aerial 3 is also coupled to a first input 5 of a multiplicative mixer 6 via the circulator 4, a second input 7 of the mixer being fed with a sample of the RF signal currently being transmitted via a coupler 8. The frequency control input 9 of oscillator 1 is fed from the output 10 of a ramp voltage generator 11 which has a reset input 12.Generator 11 may comprise in known manner an integrator circuit in the form of a high-gain inverting voltage amplifier having a capacitor connected between its output and its input, having its input connected to a reference voltage point via a series resistor, and having a controllable switch connected in parallel with the capacitor so that closure of the switch results in resetting of the output ramp. The control input of the switch (which may take the form of a field effect transistor) is fed, via the input 12, with control pulses from an output 13 of a control pulse generator 14, so that a sawtooth waveform appears at the output 10 of generator 11, resulting in the output signal of oscillator 1 exhibiting a succession of frequency sweeps in a given direction from a base value of e.g. 1OGHz.

The output 15 of mixer 6 is coupled, in practice normally via an IF amplifier (not shown), to the input 16 of an analog-to-digital converter 17 the clock or convert control signal input 18 of which is fed from an output 19 of generator 14. The output 20 of converter 17 is connected to the (multibit) serial input 21 of a shift register 22 which constitutes an input buffer for a Discrete Fourier Transform calculating circuit 23, the parallel outputs 24 of buffer 22 being coupled to corresponding inputs of circuit 23. A clock input 25 of buffer 22, and an activation signal input 26 of circuit 23, are fed from outputs 27 and 28 respectively of control pulse generator 14.

The apparatus of Figure 1 as described so far is conventional: the succession of frequency-swept segments of RF signal generated by oscillator 1 are transmitted by the aerial 3 and, if they are reflected back to the aerial 3 by a target, are mixed with a sample of the currently-transmitted signal in mixer 6 to yield a beat signal the frequency of which has components due to the range and to the relative velocity (if any) of the target. This beat signal is A/D converted in converter 17 and, for each frequency sweep of the transmitted RF, is analysed as to its spectral contents by means of the Discrete Fourier Transform (DFT) calculating circuit 23.Circuit 23 operates as a bank of n filters (only n/2 of which are distinct due to the fact that, in the present embodiment, only a single mixer 6 is employed, rather than a quadrature pair) each having a bandwidth of 1/nt where n is the number of samples used to calculate a given DFT and t is the sample period, i.e. the period of the clock pulses applied to A/D converter 17 and buffer 22.

Thus a signal appears at those of the n outputs 29 of circuit 23 which correspond to a frequency band within which a component of the beat signal outputted by mixer 6 lies, thereby indicating the range/relative-velocity combination characteristic frequency corresponding to the relevant target(s) to a resolution 1/not. The remainder of the apparatus of Figure 1, together with as yet unmentioned properties of certain components already described, are devoted to determining the velocity component of such a characteristic frequency in itself, thereby enabling the actual velocity and/or the actual range of a target to be determined.

To this end one of the n outputs 29 of the DFT calculating circuit 23 of Figure 1 is shown connected to the multibit serial data inputs 30,31 and 32 of further shift-register buffers 33,34 and 35 respectively. Buffers 33,34 and 35 comprise N1,N2 and N3 stages respectively, where N1,N2 and N3 are relatively prime, and have clock inputs 36,37 and 38 respectively fed from further outputs 39,40 and 41 respectively of control pulse generator 14. DFT calculating circuit 23 generates each time in conventional manner two quantities at each of its outputs 29, these quantities being quadrature components of the complex amplitude of the relevant spectral component. It is therefore successive samples of this complex amplitude which are fed to the buffer inputs 30,31 and 32. The N1 parallel outputs 42 of buffer 33 are connected to respective ones of the N1 inputs of a N1-point DFT calculating circuit 43, the N2 parallel outputs 44 of buffer 34 are connected to respective ones of the N2 inputs of an N2-point DFT calculating circuit 45, and the N3 parallel outputs 46 of buffer 35 are connected to respective ones of the N3 inputs of an N3-point DFT calculating circuit 47.

Activation signal inputs 48,49 and 50 of the circuits 43,45 and 47 respectively are fed from further outputs 51,52 and 53 respectively of control pulse generator 14. The N1 amplitude outputs 54 of circuit 43 are connected to respective inputs of an ambiguity reducing or resolving circuit 55, as are the N2 amplitude outputs 56 of circuit 45 and the N3 amplitude outputs 57 of circuit 47.

Circuit 55, which has an activation signal input 58 fed from a further output 59 of control pulse generator 14, has an output 60 on which appears in operation a signal indicative of the velocity component of any range/velocity combination frequency whose presence is indicated by a signal at that output 29 of DFT-calculating circuit 23 which feeds the components 33-35, as will now be explained.

The reset pulses for ramp generator 11 produced by control pulse generator 14 at its output 13 do not have a constant repetition rate R, but rather have a repetition rate which changes in steps cyclically from rate R1 to rate R2 to rate R3 to rate R1 etc, where the ratios R1:R2:R3 are reducible to the ratios N1:N2:N3, and at least N1 pulses occur each time the rate is R1, at least N2 pulses occur each time the rate is R2, and at least N3 pulses occur each time the rate is R3.

Figure 2a illustrates one complete cycle of these pulses for N1=5, N2=6 and N3=7 respectively, exactly 5, 6 and 7 of these pulses occurring at respective rates which are in this ratio, so that first five equally-spaced pulses occur in a time interval T, then six equally-spaced pulses occur in the next (equal) time interval T, and then seven equally-spaced pulses occur in the next (equal) time interval T. (After this the cycle repeats, but this is not shown in Figure 2a). In consequence the output signal of ramp voltage generator 11 is a sawtooth waveform the frequency of which varies in steps and cyclically between values R1,R2 and R3 respectively.Thus the output signal of voltage-controlled oscillator 1 comprises a succession of segments during each of which its frequency increases (or decreases) substantially linearly from a base value fo of, for example, 1OGHz, as shown in Figure 2b, the frequency returning to fO at the end of each segment when the output signal of ramp voltage generator 11 is reset. The rate of increase of frequency during each segment is the same (being determined by the time constant of the integrator circuit included in ramp voltage generator 11) and during (part of) each segment the frequency sweeps from a first given value f1 to a second given value f2.The repetition rate of the segments takes on in succession a plurality of values the ratios between which are reducible to the ratios between relatively prime integers; in the present example three such values are used in the ratios 5:6:7, for example 2.5kHz, 3.0kHz and 3.5kHz respectively.

During the generation and transmission of part of each segment, more particularly the part during which the frequency sweeps from f1 to f2, control pulse generator 14 generates a sequence of clock pulses for the A/D converter 17 and buffer 22 on its outputs 19 and 27 respectively. The intervals during which these clock pulses are generated are illustrated in Figure 2c, and during each such interval exactly n clock pulses are generated on each output 19 and 27, where n has the meaning assigned to it hereinbefore, i.e. is equal to the capacity of buffer 22 and the number of points in each transform calculated by the DFT calculating circuit 23. In practice n may be equal, for example, to 128.Thus during each sweep of the output frequency of oscillator 1 from f1 to f2 n samples of the beat signal appearing at the output 15 of mixer 6 are taken and clocked into the buffer 22. Each time buffer 22 has been loaded with n samples in this way control pulse generator 14 generates an activation or enable pulse for DFT-calculating circuit 23 on its output 28, as shown in Figure 2d, with the result that circuit 23 calculates an n-point DFT of the contents of the buffer 22 and produces the result on its (n) outputs 29. Thus a non-zero complex amplitude term (in the form of two quadrature components thereof) appears on any output 29 which corresponds to a frequency range within which a component of the beat signal appearing at the output of mixer 6 lies.It will be assumed for the purposes of the following description that the output 29 which is connected to the inputs 30-32 of the buffers 33-35 is one such output.

Each time a new n-point DFT has been calculated by the circuit 23 control pulse generator 14 generates a clock pulse on one of its outputs 39-41 and hence clocks one of the buffers 33-35. Which buffer is clocked is determined by whether the DFT corresponds to a transmitted segment of repetition rate R1, to a transmitted segment of repetition rate R2, or to a transmitted segment of repetition rate R3, the buffers clocked in the three cases being buffers 33,34 and 35 respectively. The clock pulses for these buffers generated on the outputs 39,40 and 41 respectively of generator 14 are shown in Figures 2e,2g, and 2i respectively.Thus for each of the N1=5 n-point DFTs corresponding to transmitted segments of repetition rate R1 buffer 33 is clocked, for each of the N2=6 n-point DFTs corresponding to transmitted segments of repetition rate R2 buffer 34 is clocked, and for each of the N3=7 n-point DFTs corresponding to transmitted segments of repetition rate R3 buffer 35 is clocked. Each time buffer 33 has been filled with five complex amplitude samples in this way generator 14 generates an activation signal for the five-point DFT calculating circuit 43 on its output 51, as shown in Figure 2f, with the result that a signal appears on that output 54 of circuit 43 which corresponds to a frequency range within which the frequency of any signal which has given rise to changing phase information in the samples applied to buffer 33 lies, possibly after aliassing down by the repetition rate R1 of these samples.

Similarly, each time buffer 34 has been filled with six complex amplitude samples in this way generator 14 generates an activation signal for the six-point DFT calculating circuit 45 on its output 52, as shown in Figure 2h, with the result that a signal appears on that output 56 of circuit 45 which corresponds to a frequency range within which the frequency of any signal which has given rise to changing phase information in the samples applied to buffer 34 lies, possibly after aliassing down by the repetition rate R2 of these samples.Similarly, each time buffer 35 has been filled with seven complex amplitude samples in this way generator 14 generates an activation signal for the seven-point DFT calculating circuit 47 on its output 53, as shown in Figure 2j, with the result that a signal appears on that output 57 of circuit 47 which corresponds to a frequency range within which the frequency of any signal which has given rise to changing phase information in the samples applied to buffer 35 lies, possibly after aliassing down by the repetition rate R3 of these samples.

The possible aliassing down of the frequencies indicated by the DFT-calculating circuits 43,45 and 47 means that each of these frequencies taken alone is ambiguous, and the purpose of the circuit 55 is to reduce or resolve such ambiguities, as will now be explained.

The frequency resolutions at the outputs 54,56 and 57 of DFT-calculating circuits 43,45 and 47 are R1/N1, R2/N2 and R3/N3 respectively and, because the ratios R1:Rz:R3 are reducible to the ratios N1:N2:N3, these resolutions are all the same. For example, if R1,R2 and R3 are equal to 2.5kHz, 3.0kHz and 3.5kHz respectively, and N1,N2 and N3 are equal to 5,6 and 7 respectively, then each of the successive outputs 54,56 and 57 corresponds to a frequency increment of 500Hz, the five outputs 54 together covering the frequency range 0-2.5kHz in 500Hz steps, the six outputs 56 together covering the frequency range 0-3kHz in 500Hz steps, and the seven outputs 57 together covering the frequency range 0-3.5kHz in 500Hz steps.If the signal component which gives rise to the changing phase samples applied to the buffers 33-35 has a frequency in the range 0-2.5kHz then a signal will appear at that output 54, that output 56, and that output 57 which corresponds to this frequency. If on the other hand the signal frequency is higher than 2.5kHz it will be aliassed down into the range 0-2.5kHz as far as the DFT-calculating circuit 43 is concerned, similar comments applying to the DFT-calculating circuits 45 and 47 if the signal frequency is greater than 3kHz and 3.5kHz respectively. However the actual value of the signal frequency can still be determined in such a case by taking the output signals of the circuits 43,45 and 47 together1 provided that the signal frequency is, in the present example, not greater than 105kHz (5 x 6 x 7 x 500Hz).This is because, in such a case, the relevant frequency will give rise to a unique set of three outputs, one from each of the circuits 43,45 and 47. The circuit 55 in effect decodes each possible such set (there being 5 x 6 x 7 = 210 possibilities in the present example) and generates a signal indicative of the actual signal frequency (to a resolution of 500Hz in the present case) on its output 60.To this end circuit 55 may make use of the Chinese Remainder Theorem which states that the number x represented in the relatively prime modulus set N1,N2 Nk by the set of residues x1,x2,....xk is given by the summation of the weighted residues, modulo N: x = (Clxl+C2x2+ .... + Ckxk) modulo N where N = (N1)(N2)----(Nk) and Ci = [N/Ni].([N/Ni]-1 modulo Ni) (i = 1,2 k) (The coefficients Ci are in fact the numbers which are unity in the respective modulus and zero in all the others). Thus, for example, the circuit 55 may comprise a set of multipliers for weighting the various xi (i=1,2,...k), an addition unit to calculate the sum, and a modulo-N unit such as a divider-by-N to bring the result back within range.In the present example N1=5, N2=69 N3=7 and N=210.

Thus C1 = 42.42-1 modulo 5 = 42x3 = 126.

C2 = 35.35-1 modulo 6 = 35x5 = 175 and C3 = 30.30-1 modulo 7 = 30x4 = 120.

Therefore x = (126 x1 + 175 x2 + 120 x3) modulo 210.

If, for example, the signal component which gives rise to the changing phase samples applied to the buffers 33-35 has a frequency of +22kHz (this in fact being approximately the Doppler shift at 1OGHz produced by a relative target velocity towards the radar apparatus of Mach 1), this will be aliassed down to 2kHz as far as the DFT calculating circuit 43 is concerned, to 1kHz as far as the DFT calculating circuit 45 is concerned, and to 1kHz as far as the DFT calculating circuit 47 is concerned. Thus a signal will appear on the fourth of the five outputs of circuit 43, on the second of the six outputs of circuit 45 and on the second of the seven outputs of circuit 47. Applying the Chinese Remainder Theorem: x = (126x4+175x2+120x2) modulo 210 = 1094 modulo 210 = 44.

Thus the input frequency is 44 x 500Hz = +22kHz as required.

This frequency fd can then be used to calculate the relative velocity V of the target from the well-known formula V=fdc/2fa, where c is the velocity of light and fa is the mean effective output carrier frequency of the radar ((frf)/2) in Figure 2b). Alternatively or in addition the frequency fd can be subtracted from the mean frequency to which the relevant output 29 of DFT-calculating circuit 23 corresponds, to give a frequency fr which is representative of the actual range R of the target according to the well-known formula R=cfr/2oc, where a is the rate of change of the carrier frequency of the radar during each frequency sweep.

If the target giving rise to the 22kHz doppler frequency shift had been receding from the radar apparatus instead of approaching it, this frequency shift would have been negative. Such a negative frequency would have been aliassed to 500Hz (2.5kHz-22kHz modulo 2.5kHz) as far as the DFT calculating circuit 43 is concerned, to 2kHz (3.0kHz-22kHz modulo 3kHz) as far as the DFT-calculating circuit 45 is concerned, and to 2.5kHz (3.5kHz-22kHz modulo 3.5kHz) as far as the DFT-calculating circuit 47 is concerned. Thus a signal would have appeared on the first of the five outputs of circuit 43, on the fourth of the six outputs of circuit 45, and on the fifth of the seven outputs of circuit 47. Applying the Chinese Remainder Theorem: x = (126 x 1 + 175 x 4 + 120 x 5) modulo 210 = 1426 modulo 210 = 166.

Thus the input frequency would have been calculated as 166 x 500Hz = 83kHz, which is the frequency 105kHz-22kHz modulo 105kHz to which a frequency of -22kHz is aliassed by a sampling process at 105kHz (5 x 6 x 7 x 500Hz) as required.

As an alternative circuit 55 may be constituted, for example, by a suitably programmed look-up table.

Control pulse generator 14 generates an activation or enable signal for the circuit 55 on its output 59 each time a new set of DFTs has been calculated by the circuits 43,45 and 47, as shown in Figure 2k.

The repetition rates R1,R2 and R3 in the above example are chosen on the basis that the relative velocities of the targets are never greater than those which give rise to positive or negative doppler frequency shifts of 105/2kHz, so no ambiguity arises between positive and negative relative velocities. If higher velocities were to occur the result of using the Chinese Remainder Theorem or another form of decoder would have still been ambiguous, but obviously this ambiguity would not be as great as that present at the outputs of any of the circuits 43,45 or 47 when taken individually.In principle as large a range of possible input frequencies as desired can be accommodated unambiguously by increasing the values of N1 and/or N2 and/or N3, and/or by providing more than three segment repetition rates Ri and corresponding, relatively prime, Ni, the latter entailing the provision of more buffers and DFT-calculating circuits similar to the buffers 33-35 and circuits 43,45 and 47. Conversely the use of just two repetition rates R1 and R2 and corresponding relatively prime N1 and N2 may suffice in some cases to unambiguously accommodate the possible range of input frequencies.

It will be appreciated that, as described, only the Doppler frequency shift of a target which results in a beat frequency corresponding to that output 29 of circuit 23 which is connected to the buffers 33-35 is determined. In practice it will usually be the case that the Doppler frequency shift of any target within range be determinable and, to this end, the buffer inputs 30-32 may be made connectable to any of the left-hand n/2 or any of the right-hand n/2 of the outputs 29 at will, either manually or, for example, automatically in response to a signal amplitude of at least a predetermined value being indicated at the relevant output. Alternatively the system of components 33,34,35,43,45,47 and 55 may be duplicated a number of times so that one is present corresponding to, and fed from, each of the n/2 outputs 29 which produce independent output signals.

On the other hand sometimes, e.g. in an altimeter application, it may be the case that only a single beat frequency is of interest. If this is so the buffer 22 and DFT-calculating circuit 23 may be replaced by a pair of quadrature mixers fed with the beat frequency of interest as a reference.

It will be evident that, although it is most efficient to transmit exactly N1,N2 and N3 segments at the repetition rates R1,R2 and R3 respectively during each cycle of these three repetition rates, this is not essential. It is merely necessary that at least N1,N2 and N3 segments respectively be transmitted during each cycle, enabling the buffers 33,34 and 35 to be completely filled each time.

It will furthermore be evident that it is not essential that the various Ni be relatively prime, merely that the ratios therebetween be reducible to the same quotients of relatively prime integers as those to which the ratios between the corresponding Ri are reducible. Thus, for example, alternative values for N1,N2 and N3 in the embodiment described are ten, twelve and fourteen respectively.

The control pulse signal generator 14 may comprise a clocked multistage counter selected stages of which are connected to various decoders which are constructed to generate output signals corresponding to respective ones of those shown in Figures 2a and 2c-2k. To facilitate this the clocking rate of the counter is preferably equal to, or a multiple of, the clocking rate of the A/D converter 17 and buffer 22, and also a multiple of the product of the repetition rates R1,R2 and R3.

Although as shown in Figure 1 the same aerial 3 is used to both transmit the output signal of oscillator 1 and receive the reflected segments for application to the mixer 6, it will be appreciated that this is not essential. Respective aerials may be used for transmission and reception, if desired.

Although as described the successive segments of RF signal generated by the oscillator 1 of Figure 1 and transmitted are contiguous - see Figure 2b - it will be appreciated that this too is not essential. For example, each of the segments at the relative repetition rates five and six may be terminated when the transmitted frequency reaches the maximum frequency transmitted within a segment at the relative repetition rate seven.

From reading the present disclosure, other modifications will be apparent to persons skilled in the art. Such modifications may involve other features which are already known in the design, manufacture and use of radar apparatuses and component parts thereof and which may be used instead of or in addition to features already described herein. Although claims have been formulated in this application to particular combinations of features, it should be understood that the scope of the disclosure of the present application also includes any novel feature or any novel combination of features disclosed herein either explicitly or implicitly or any generalisation thereof, whether or not it relates to the same invention as presently claimed in any claim and whether or not it mitigates any or all of the same technical problems as does the present invention The applicants hereby give notice that new claims may be formulated to such features and/or combinations of such features during the prosecution of the present application or of any further application derived therefrom.

Claims (5)

CLAIM(S)

1. A method of resolving range/doppler coupling in a swept-frequency radar apparatus, in which method the apparatus transmits a succession of segments of radio frequency signal during each of which the frequency of the signal sweeps from a first given value to a second given value at a rate which is the same for each sweep, the transmitted segments, after reflection by said target back to the apparatus, are each mixed with a sample of the signal presently transmitted to yield a beat signal the frequency of which has components due to the range and to the relative velocity (if any) respectively of the target, and the component due to the relative velocity is determined from the beat signal obtained from a plurality of the reflected segments, characterized in that the repetition rate of said segments takes on in succession first and second values R1 and R2 respectively the ratio between which is reducible to the quotient of first and second integers which are relatively prime, the numbers of segments transmitted at the repetition rates R1 and R2 are at least N1 and N2 respectively where the ratio N1:N2 is equal to or reducible to said quotient, for each said repetition rate Ri an Ni-point Discrete Fourier Transform is taken of the complex amplitudes of the beat signals resulting from Ni successive reflected segments to yield a respective frequency which is a possibly ambiguous representation of said component due to the relative velocity, and the possible ambiguity is reduced or resolved by taking said respective frequencies together.

2. A method as claimed in Claim 1, wherein the beat signal obtained from each reflected segment is analysed in respect of frequency and phase by taking a Discrete Fourier Transform thereof, and each Ni-point Discrete Fourier Transform is taken of the Ni complex amplitudes appearing at corresponding output points of the Discrete Fourier Transforms taken of the beat signals resulting from successive reflected segments at the repetition rate Ri.

3. Radar apparatus comprising a swept-frequency RF oscillator the output of which is coupled to an aerial, which oscillator is constructed to generate for transmission a succession of segments of radio-frequency signal during each of which the frequency of the signal sweeps from a first given value to a second given value at a rate which is the same for each sweep, a mixer having inputs coupled to the output of the oscillator and to an aerial respectively, for mixing each transmitted segment, after reflection by a target, with a sample of the signal presently transmitted to yield a beat signal the frequency of which has components due to the range and to the relative velocity (if any) respectively of the target and means for determining the component due to the velocity (if any) from the beat signal obtained from a plurality of the reflected segments, characterized in that the oscillator is constructed to generate for transmission segments the repetition rate of which takes on in succession first and second values R1 and R2 respectively the ratio between which is reducible to the quotient of first and second integers which are relatively prime and in such manner that the numbers of segments transmitted at the repetition rates R1 and R2 are at least N1 and N2 respectively where the ratio N1: :N2 is equal to or reducible to said quotient, and in that said means comprises means for taking, for each said repetition rate Ri, an Ni-point Discrete Fourier Transform of the complex amplitudes of the beat signals resulting from Ni successive reflected segments to yield a respective frequency which is a possibly ambiguous representation of said component due to the ralative velocity, and for reducing or resolving the ambiguity by taking said respective frequencies together.

4. A method of determining the relative velocity between a target and a radar apparatus, substantially as described herein with reference to the drawings.

5. Radar apparatus substantially as described herein with reference to the drawings.