WO2014123433A1

WO2014123433A1 - A method of and a circuit for radar signal compression

Info

Publication number: WO2014123433A1
Application number: PCT/PL2013/000021
Authority: WO
Inventors: Aleksander ŁABUDZIŃSKI; Mariusz Zych
Original assignee: BUMAR ELEKTRONIBUMAR Spółka Akcyjna
Priority date: 2013-02-07
Filing date: 2013-02-15
Publication date: 2014-08-14
Also published as: PL402704A1; PL222895B1

Abstract

A method of and a circuit for radar signal compression according to the invention can be applied in radar systems, especially those generating and processing broadband radar pulses with constant amplitude and linear or nonlinear frequency modulation. In a composite radiolocation pulse composed of a definite number of subpulses with different carrier frequencies, signal compression can be accomplished with the use of analytical properties of signals matched to subpulses of the composite radar signal and storing them in a non-volatile memory block. Further, determination of an optimum low-pass representation of radar subpulses includes a stage of analog-digital conversion and a matched filtering stage in a bank of matched filters. Operation of each matched filter consists in calculation of two convolutions of the composite radar signal with analytic signals matched to subpulses.

Description

A METHOD OF AND A CIRCUIT FOR RADAR SIGNAL COMPRESSION

The subject of the present invention is a method of and a circuit for compression of radar signal to be applied in radar devices, especially those in which broadband pulses with constant amplitude and linear or nonlinear intra-pulse frequency modulation are generated and processed.

There is a well-known method of achieving a desired correlation properties of the radar uncertainty function in devices utilizing pulsed signals where the generated pulses must be modulated on the sending side and subject to compression in the time domain on the receiving side. Usually, angle modulation is utilized which guarantees good use of energy properties of the sending systems. Moreover, a pulsed radiolocation signal is composed of a number of subpulses with different carrier frequencies ensuring separation of subpulses. In order to obtain a low-pass representation of the signal, quadrature demodulation circuits are commonly used at outputs of which the in-phase component and the quadrature component are obtained. As a result, the signal is represented in complex space. The complex low-pass representation determined this way is subject to optimum filtering maximizing the signal power to noise and interference power ratio at the matched filter output.

Description of patent No. PL211553 reveals a matched filtering method characterized in that the radiolocation echo signal transferred to the base band at the intermediate-frequency is subject to analog-digital conversion with non-uniform sampling cycle in an analog-to-digital converter, whereas subsequent samples are taken in instants of time in which the sounding signal phase increases by π radians, and then are added in a set of N accumulators. The matched filtering circuit according to the invention is distinctive in that it comprises an analog-to-digital converter connected to input of the set of N accumulators. Outputs of each of N accumulators are connected to demultiplexer output of which represents response of the matched filter.

Known is a radiolocation signal compression circuit described in publication entitled 'Radar handbook. Second edition' by M. Skolnik published by McGraw-Hill Inc. in 1990 where a radar receiving channel is described comprising an analog-to-digital converter, a quadrature demodulator, a low-pass filter, and a set of four matched filters accomplishing compression of the signal. The analog-to-digital converter is used to represent the analog signal in digital form. Output signals of the quadrature demodulator accomplishing a low-pass representation of the signal are subject to low-pass filtering thus eliminating undesirable products of frequency conversion. Matched filtering is accomplished as a convolution of complex signals representing the echo and a standard. This requires four real convolutions of component signals to be calculated and two summation operations performed.

The objective of the present invention is to provide a method of and a circuit for radar signal compression characterized with reduced number of calculation operations in the area of digital processing of signals and simplified design, in which the number of functional circuits included in the signal compression circuit will be less compared to the prior art.

The method of compression of radar signal composed of N subpulses with different carrier frequencies and separated from each other in the frequency domain according to the presents invention is characterized in that the composite radar signal generated in the standard composite radar signal generator is subject to conversion in an analog-to-digital converter with uniform sampling and multiplied in the multiplier with tabulated samples of signal described by means of a time window function, preferably the Hamming window. Result of the multiplication is subject to filtering in the Hilbert filter. Imaginary component of the standard analytic signal from the Hilbert filter is inverted in the inverter and transmitted to the first demultiplexer. Real component of the standard analytic signal from the Hilbert filter output is transmitted directly to the second demultiplexer. Both demultiplexers are controlled by means of the subpulse selection signal within structure of the composite radar signal. Real and imaginary components of the standard analytic signal are stored in a block of non-volatile memory with separation into real and imaginary components retained for each subpulse of the composite radar signal. Next, filtering in a bank of matched filters is accomplished by means of multiplying radar echo signal samples by samples of real part and imaginary part of the analytic standard signal in separate multiplying circuits for each of N echo signal subpulses separately. Output signals in N processing channels constitute real component and imaginary component of the signal after compression for N echo signal subpulses.

The method is characterized in that matched filtering of a radar signal can be carried out both in the time domain and the frequency domain.

The radar signal compression circuit according to the presents invention that comprises an analog-to-digital converter, an inverter, a Hilbert filter, a signals multiplying circuits, a memory block, a demultiplexer of signals in time domain, and a bank of matched filters is characterized in that the address counter in which the radiolocation signal generator is synchronized by means of a trigger pulse and controlled by means of the subpulse selection signal, has output coupled with input of the first analog-to-digital converter. The standard composite radar signal generator has output coupled with input of the first analog-to-digital converter output of which is connected to the first input of the signal multiplication circuit, and second input of which is connected to output of the first block of non-volatile memory. Input of the non-volatile memory is connected to output of the address counter, whereas the first input of the address counter is connected to the clock signal, and the second input is connected to the triggering signal. Output of the signal multiplier is coupled with input of the Hilbert filter the first output of which is connected to input of the inverter. Output of the inverter is connected to input of the first demultiplexer control input of which is connected to the subpulse selection signal. The second output of the Hilbert filter is connected to input of the second demultiplexer control input of which is connected to the subpulse selection signal. All outputs of the first demultiplexer are coupled with non-volatile memory block cells storing values of imaginary component of the analytic matched signal. All outputs of the second demultiplexer are connected to non-volatile memory block cells storing values of real components of the analytic matched signal. The intermediate-frequency radar signal is coupled with input of an analog-to-digital converter output of which is connected to the bank of matched filters, whereas the second inputs of each of the filters from the matched filters bank are connected to outputs of the second non-volatile memory block. Input of the second non-volatile memory block is connected to output of the second address counter where the same control signals are supplied to control input of the first address counter and to control input of the second address counter. Outputs of each of the filters of the bank of matched filters constitute real and imaginary component of the complex radar signal after compression.

A favorable feature of the method and the circuit according to the presents invention consists in that by using properties of the Hilbert space one obtains significant reduction of calculation effort related to converting a signal to an optimum low-pass form according to the criterion of maximization of the signal power to noise and interference power ratio at the compression filter output.

Application of the invention reduces the number of sub-circuits included in the signal compression circuit with respect to the prior art. Compared to the known signal processing methods, eliminated are such circuits as quadrature denominators, low-pass filters, and partial convolution adders.

The method and the circuit according to the present invention is favorably characterized in that matched filtering is reduced to calculation of two linear convolutions of a real signal with Hilbertian standard. Matched filtering methods utilized to date require calculation of four partial convolutions of real and imaginary components of echo and standard signals and performing two adding operations.

Subject-matter of the present invention is explained by means of its embodiment shown in figures, of which Fig. 1 presents a schematic block diagram showing processing of a composite radar signal comprising, in general, N subpulses in which the method of radar signal matched filtering in time domain was applied; Fig. 2 presents the matched filtering circuit; Fig. 3 presents frequency spectrum of actual radiolocation signal composed of two subpulses with linear frequency modulation; Fig. 4 shows spectrum of analytic signal matched to the first subpulse modulated by means of Hamming time window function; Fig. 5 presents spectrum of analytic signal matched to the second subpulse modulated by means of Hamming time window function; Fig. 6 presents convolution of the first subpulse signal with real component of the analytic matched signal in linear scale; Fig. 7 presents convolution of the first subpulse signal with imaginary component of the analytic matched signal in linear scale; Fig. 8 presents convolution of the first subpulse signal with real component of the analytic matched signal in logarithmic scale; Fig. 9 presents convolution of the first subpulse signal with imaginary component of the analytic matched signal in logarithmic scale; Fig. 10 presents logarithmic plot of the matched filter output signal obtained by means of the method consisting in calculation of four partial convolutions and two operations of adding the complex radar signal to the matched signal known in the prior art; and Fig. 11 presents logarithmic plot of output of the filter matched to the first subpulse obtained by means of modification according to the present invention with the use of preferable properties of signals represented in Hilbert space obtained as a result of calculation of two partial convolutions.

The method according to the present invention in an example embodiment includes the stage in which the standard composite radar signal generator RSG, comprising at least two control inputs, one for subpulse selection NS and another for triggering signal PR1 , generates the intermediate-frequency radar signal composed of two subpulses. Subpulses making the composite radar signal have duration of 120 μβ and frequency deviation of 4 MHz. The intermediate frequency of the first subpulse is 63 MHz, and the intermediate frequency of the second subpulse is 77 MHz. In the subsequent stage, the composite radar signal is subject to analog-digital conversion in the first converter ADO with the use of band-pass sampling method with sampling frequency /s equaling 100 MHz. As a result, subpulse signal spectra are being shifted, and subpulse carrier frequencies are 23 MHz (f_s - 77 MHz) and 37 MHz (fs - 63 MHz). In the next stage, the discrete composite intermediate-frequency radar signal is multiplied in the multiplier by the time window signal. Coefficients of the Hamming time window function are stored in the first non-volatile memory block. Moreover, reading of Hamming time window coefficients from the first memory block is controlled by means of the first address counter LI synchronized by means of clock signal CLK and trigger signal PR1. Further, the modulated signal is in the next stage subject to filtering in Hilbert filter HILB, however the HILB filter output corresponding to imaginary component of the analytical signal is inverted. Frequency spectrum of the first subpulse analytic signal is shown in Fig. 4. Frequency spectrum of the second subpulse analytic signal is shown Fig. 5. Output signals of demultiplexers DMUX1, DMUX2 controlled by means of the subpulse selection signal NS are in the next stage stored in memory cells ROM_1R, ROMn and ROM^, ROM₂i, respectively, in the second non-volatile memory block controlled by means of the second address counter L2. The signals are hi_R, hn and h_2R, h₂i analytic signals modulated by means of the time window function and matched to the standard composite radar signal generated in the first stage. Further, the composite intermediate-frequency radar echo signal A_SYG is subject to analog-digital conversion in the second converter ADC2 with the use of band-pass sampling method with sampling frequency s equaling 100 MHz. The composite radar echo signal is composed of two subpulses (N = 2) with linear frequency modulation, duration period of 120 μβ, and frequency deviation 4 MHz for carrier frequencies equaling 77 MHz and 63 MHz. As a result, subpulse spectra are being shifted as shown in Fig. 3 and Fig. 4, and subpulse carrier frequencies are 23 MHz and (f_s - 77 MHz) and 37 MHz (f_s - 63 MHz), respectively. However, the presented parameters by no means limit the possibility to utilize the present invention in the case of any other number of subpulses, other modulation types including nonlinear modulation methods, or for other carrier frequencies.

Γη the next stage, matched filtering is determined which utilizes favorable properties of the signal representation in Hilbert space obtained as a result of calculation of two partial convolutions for each of the two subpulses. Plots of partial convolutions REtYi] and IM[Yi] of the filter matched to the first subpulse in linear scale are presented in Fig. 8, and in logarithmic scale in Fig. 9. Plot of modulus of the output signal from filter matched this way is presented in Fig. 10. The method of matched signal filtering accomplished in Hilbert space according to the invention consists in calculating two partial convolutions of real component of the signal with analytical standard signal. Plot of the output signal modulus of the filter matched to the first subpulse accomplished according to the invention is presented in Fig. 11. Results of compression presented in Fig. 10 and Fig. 11 are identical. In view of the above, accomplishment of matched filtering according to the invention is equivalent to methods utilized to date. The number of partial convolutions is being reduced from four to two; moreover, adding operations are eliminated. Further, accomplishment of matched filtering according to the invention allows to remove phase demodulator together with low-pass filters following it from the radar's receiving channel.

The circuit according to the invention in the presented example embodiment processes a composite radar signal composed of two subpulses (N = 2). The circuit comprises a standard composite radar signal generator 1 that is synchronized by means of trigger pulse PR1 and controlled by means of the subpulse selection signal NS. Output of the standard composite radar signal generator 1 is connected to input of analog-to-digital converter 2. The signal is subject to analog-digital conversion with the use of band-pass sampling method with sampling frequency f_s. Output of converter 2 is coupled with the first input of the signal multiplication circuit 3 second input of which is connected to output of the first non-volatile memory block 5 control input of which is connected to output of address counter 4 the first control input of which is connected to clock signal CLK, while the second control input is coupled with the triggering source PR1 that is also supplied to the standard composite radar signal generator block RSG. Output of the signal multiplication circuit 3 is connected to input of Hilbert filter 6 first output of which is connected to inverter 7, while the second output is connected to input of demultiplexer 9 controlled by means of the subpulse selection signal NS. Output of inverter 7 is connected to input of demultiplexer 8 controlled by means of the subpulse selection signal NS. Outputs of demultiplexer 8 number of which depends on the number of subpulses Ij, I₂, I_N are connected to inputs of memory cells ROMn, ROM₂₁, ROMN! of memory block 10. Outputs of demultiplexer 9 number of which depends on the number of subpulses Ri, R₂, RN are connected to inputs of memory cells ROMIR, ROM R, ROMNR of memory block 10. Outputs of memory cells ROMn, ROM₂i, ROMNI and ROMi , ROM₂R, ... , ROMN_R of memory block 10 are connected to inputs of N matched filters of the filter bank 12. Further, control input of memory block 10 is connected to output of address counter 13 the first control input of which is connected to clock signal CLK and the second input is connected to triggering signal PR1. First inputs of N matched filters from filter bank 12 are connected to output of analog-to-digital converter 11 utilizing the band-pass sampling method with sampling frequency /s, whereas input of converter 11 constitutes the intermediate frequency radar echo signal A_SY_G. In the case of processing composite radar signals composed of more than two subpulses (N > 2), the circuit is subject to appropriate logical and electrical modification without any limitation of possibility to apply it according to the above-described example embodiment.

Claims

Patent claims

1. A method of compression of a radar signal composed of N subpulses with different carrier frequencies and separated from each other in the frequency domain characterized in that the composite radar signal generated in the standard composite radar signal generator (1) is subject to conversion in analog-to-digital converter (2) with uniform sampling and multiplied in multiplier (3) by tabulated samples of signal described by means of a time window function, preferably Hamming window, and result of the multiplication is subject to filtering in Hilbert filter (6) where imaginary component of standard analytic signal from the Hilbert filter is inverted in inverter (7) and transmitted to first demultiplexer (8), and the real component of the standard analytic signal from output of the Hilbert filter is transmitted directly to second demultiplexer (9), while both demultiplexers are controlled by means of the subpulse selection signal within the composite radar signal structure; next, real and imaginary components of the standard analytic signal are stored in non- volatile memory block (10) with separation into real and imaginary components retained for each subpulse of the composite radar signal; then, filtering in bank (12) of matched filters is accomplished by multiplying, in separate multipliers, radar echo signal subpulses by samples of real part and imaginary part of the analytic standard signal for each of N echo signal subpulses separately, whereas input signals in N processing channels constitute real part and imaginary part of the signal after compression for N echo signal subpulses.

2. The method according to claim 1 characterized in that matched filtering of radar signal can be accomplished both in the time domain and the frequency domain.

3. A radar signal compression circuit comprising an analog-to-digital converter, an inverter, a Hilbert filter, a signal multiplier circuit, a memory block, a circuit for signal demultiplexing in the time domain, and a bank of matched filters characterized in that the address counter in which the standard radar signal generator (1), synchronized by means of a trigger pulse and controlled by means of the subpulse selection signal, has output coupled with input of analog-to-digital converter (2) and standard radar signal generator (1) has output coupled with input of analog-to-digital converter (2) output of which is connected to input of signal multiplier (3) the second input of which is connected to output of non-volatile memory block (5) input of which is connected to output of address counter (4), whereas the first input of address counter (4) is connected to clock signal, and the second input is connected to triggering signal; further, output of signal multiplier (3) is connected to input of Hilbert filter (6) first output of which is connected to input of inverter (7), whereas output of inverter (7) is connected to input of demultiplexer (8) control input of which is connected to the subpulse selection signal, while the second output of Hilbert filter (6) is connected to input of demultiplexer (9) control input of which is connected to the subpulse selection signal, while all outputs of demultiplexer (8) are connected to cells of non-volatile memory block (10) storing values of imaginary component of the analytic matched signal, while all outputs of demultiplexer (9) are coupled with cells of non-volatile memory block (10) storing values of real component of the analytic matched signal; moreover, the intermediate frequency radar signal is coupled with input of analog-to-digital converter (HJ output of which is connected to bank (12) of matched filters, whereas second inputs of each of matched filters of bank (12) are connected to outputs of non- volatile memory block (10) input of which is connected to output of address counter (13) where the same control signals are supplied to the control input of address counter (5) and to control input of address counter (13), whereas outputs of each of the filters of bank (12) of matched filters constitute real and imaginary components of the complex compressed radar signal.