A radar signal processing method based on non-equispaced modulation DDMA-MIMO

Abstract

The present invention provides a radar signal processing method based on non-equispaced modulation DDMA-MIMO. On the one hand, since non-coherent accumulation results of Q sub-RD maps each with a size of 1 / Q RD map are respectively subjected to threshold detection during threshold detection, the detection computation amount does not increase, but due to the adoption of non-coherent accumulation, the detection performance is significantly improved. On the other hand, since the peak positions corresponding to different TXs are determined by comparing the peak amplitudes corresponding to the target positions in the Q accumulation results, and then the velocity ambiguity and angle measurement are resolved, its reliability is significantly higher than the method of amplitude comparison in the angle domain in the prior art, improving the detection performance.

Description

Technical Field

[0001] The present invention relates to the field of radar technology, and in particular, to a radar signal processing method based on non-uniformly spaced modulation DDMA-MIMO. Background Art

[0002] Vehicle-mounted millimeter-wave radar estimates information such as the distance, speed, and angle of a target by transmitting a frequency-modulated continuous wave (FMCW) and using the echo signal reflected by the target, and plays an extremely important role in the autonomous driving system.

[0003] Doppler division multiple access multiple-input multiple-output (DDMA-MIMO) radar simultaneously transmits signals through different transmit antennas (TX) and performs linear phase modulation on each transmit antenna, so that the waveforms of different transmit antennas can be directly separated in the Doppler domain. Compared with the traditional time-division MIMO radar, the DDMA-MIMO radar has stronger target detection performance and anti-interference ability, and has received extensive attention from all walks of life in recent years. In the non-uniformly spaced modulation DDMA-MIMO radar, since the signals of different transmit antennas TX are non-uniformly distributed in the Doppler dimension, the signals of different transmit antennas TX can be quickly distinguished by using the relative positions of different peaks in the Doppler dimension, overcoming the problem that it is difficult to correctly distinguish due to the equally spaced distribution of the signals of different transmit antennas TX in the spectrum in the equally spaced modulation DDMA-MIMO radar. In the non-uniformly spaced modulation DDMA-MIMO radar, the modulation phase of the signal transmitted by the m-th transmit antenna TX in the n-th pulse repetition period can be expressed as

[0004]

[0005] where m = 1, 2,..., M, Q is the number of equal divisions of the Doppler dimension during DDMA modulation, and Q is an integer greater than M, M is the number of transmit antennas, b m takes an integer between 0 and M - 1, and b m ≤Q is determined according to the modulation phase.

[0006] Since the number Q of spectrum divisions of the non-uniformly spaced modulation radar is more than the number M of transmit antennas, if non-coherent integration is directly performed on the spectra of Q different regions, the detection performance cannot reach the optimal. Summary of the Invention

[0007] The present invention proposes a radar signal processing method based on non-uniformly spaced modulation DDMA-MIMO, which can more accurately calculate the velocity ambiguity and angle measurement without increasing the amount of computation.

[0008] A radar signal processing method based on non-uniformly spaced modulation DDMA-MIMO includes the following steps:

[0009] Step S1: Perform two-dimensional Fourier transform on the echo signals of different receiving channels to obtain N range-Doppler (RD) maps corresponding to N receiving channels respectively; N is the number of receiving channels.

[0010] Step S2: Perform the first non-coherent integration on the N RD maps corresponding to N different receiving channels to obtain the RD map after the first non-coherent integration.

[0011] Step S3: According to the specific modulation method of the non-equispaced modulated DDMA-MIMO signal at the transmitting end, evenly divide the RD map after the first non-coherent integration into Q parts in the Doppler dimension to obtain Q sub-RD maps; where Q is the number of equal divisions of the Doppler dimension during DDMA modulation of the transmitted signal.

[0012] Step S4: Extract M continuously distributed sub-RD maps from the Q sub-RD maps to form a sub-RD map combination, and a total of Q different sub-RD map combinations are formed.

[0013] S5: For each sub-RD map combination, perform the second non-coherent integration on all M sub-RD maps to obtain the sub-RD map after non-coherent integration corresponding to each combination, and a total of Q sub-RD maps after non-coherent integration are obtained.

[0014] S6: Perform threshold detection on the sub-RD maps after non-coherent integration obtained in S5 respectively, obtain the range index Irange and Doppler index Idop_sub of the target in each sub-RD map after non-coherent integration, and determine the target range information according to the range index Irange.

[0015] Step S7: Compare the amplitudes of each target obtained in S5 in the Q sub-RD maps after non-coherent integration, select the target with the largest amplitude as the true target; determine the true speed of the target according to the serial number of the sub-RD map after non-coherent integration where the obtained true target is located and its Doppler index Idop_sub.

[0016] Step S8: Estimate the angle of the target.

[0017] Preferably, in step S4, the positions Int1, Int2, …, IntQ of the same target in each sub-RD map after non-coherent integration are the same.

[0018] Preferably, in step S7, the method for determining the true speed of the target is as follows:

[0019] Assume the target is in sub-RD map Int qIf the amplitude in it is the largest, the Doppler dimension index corresponding to the true target speed is Idop_total = Idop_sub + (q - 1)·Nfft_dop / Q, where q is an integer between 1 and Q; Nfft_dop is the number of FFT points in the Doppler dimension; further determine the true speed of the target according to Idop_total.

[0020] Preferably, in step S8, the method for estimating the angle of the target is:

[0021] Step S81: Combine the phase modulation methods of the transmission signals of different transmission antennas TX during signal transmission, and determine the Doppler indexes corresponding to each transmission antenna TX1 to TXM according to the true Doppler index Idop_total of the target;

[0022] Step S82: According to the target range index I range and the Doppler indexes corresponding to different transmission antennas TX, extract the data corresponding to different transmission antennas TX in the RD diagram corresponding to each receiving antenna RX, and arrange the data values corresponding to the transceiver channels formed by different receiving antennas RX and transmission antennas TX in sequence according to the antenna layout to obtain the signal of the virtual receiving antenna with a length of N·M, perform angle-domain Fourier transform on it, and detect the target in the angle to obtain the target angle information.

[0023] Preferably, in step S8, the method for determining the Doppler indexes corresponding to each transmission antenna TX1 to TXM is:

[0024] Assume that at the transmitting end, the step size of the phase modulation of the transmission antenna TXm between different PRTs is 2πb m / Q, and the Doppler index I dop_m corresponding to the transmission antenna TXm is expressed as: where N fft_dop is the number of FFT points in the Doppler dimension, m = 1, 2,..., M, and M is the number of transmission antennas TX.

[0025] Preferably, in steps S2 and S5, the non-coherent accumulation method adopts taking the amplitude average, taking the modulus square average or taking the logarithm average.

[0026] Preferably, in step S6, the method of threshold detection is the constant false alarm rate CFAR detection method.

[0027] The present invention has the following beneficial effects:

[0028] The present invention provides a radar signal processing method based on non-equispaced modulation DDMA-MIMO. On the one hand, since non-coherent accumulation results of Q sub-RD maps, each with a size of 1 / Q RD map, are respectively subjected to threshold detection during threshold detection, the detection computation amount does not increase. However, due to the use of non-coherent accumulation, the detection performance is significantly improved. On the other hand, by comparing the peak amplitudes corresponding to the target positions in the Q accumulation results to determine the peak positions corresponding to different TXs, and then resolving the velocity ambiguity and angle measurement, its reliability is significantly higher than the angle domain amplitude comparison method in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 FIG. is a schematic diagram of the processing flow of a radar signal processing method based on non-equispaced modulation DDMA-MIMO provided by the present invention.

[0030] Figure 2 FIG. is an example of the target peak distribution corresponding to TX1 to TXM in the RD maps after non-coherent accumulation of different RXs.

[0031] Figure 3 FIG. is a schematic diagram of the non-coherent accumulation results corresponding to Q different combinations of sub-RD maps. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0032] The following describes the present invention in detail with reference to the drawings and embodiments:

[0033] As Figure 1 shown in the schematic diagram of the processing flow of a radar signal processing method based on non-equispaced modulation DDMA-MIMO provided by the present invention, the method specifically includes the following steps:

[0034] Step S1: Perform two-dimensional Fourier transform (2DFFT) on the echo signals of different receiving channels (RX) to obtain N range-Doppler (RD) maps respectively corresponding to N receiving channels.

[0035] Step S2: Perform the first non-coherent accumulation on the N RD maps respectively corresponding to N different receiving channels to obtain the RD map after the first non-coherent accumulation.

[0036] Step S3: According to the specific modulation method of the non-equispaced modulation DDMA-MIMO signal at the transmitting end, evenly divide the RD map after the first non-coherent accumulation into Q parts in the Doppler dimension to obtain Q sub-RD maps, denoted as RD1, RD2,..., RD Q , where Q is the number of equal divisions of the Doppler dimension during DDMA modulation of the transmitted signal.

[0037] Step S4: Extract M consecutively distributed sub-RD maps from the Q sub-RD maps to form a sub-RD map combination. A total of Q different sub-RD map combinations can be formed, including:

[0038] The first sub-RD map combination: RD1, RD2, …, RD M ;

[0039] The second sub-RD map combination: RD2, RD3, …, RD M+1 ;

[0040] ……

[0041] The (Q - M + 1)-th sub-RD map combination: RD Q-M+1 , RD Q-M+2 , …, RD Q ;

[0042] The (Q - M + 2)-th sub-RD map combination: RD Q-M+2 , RD Q-M+3 , …, RD Q , RD1;

[0043] ……

[0044] The Q-th sub-RD map combination: RD Q , RD1, …, RD M-1 .

[0045] At this time, for the same target, the M peaks corresponding to the M transmitting antennas are respectively distributed in M of the sub-RD maps, and their positions in each sub-RD map are the same, as Figure 2 shown.

[0046] Step S5: For each sub-RD map combination, perform a second non-coherent accumulation on all M sub-RD maps to obtain the non-coherently accumulated sub-RD map corresponding to each combination. A total of Q accumulated sub-RD maps are obtained, as Figure 3 shown, and are respectively denoted as Int1, Int2, …, Int Q ;

[0047] Step S6: If the M peaks corresponding to the target are exactly distributed in the selected M sub-RD maps, then the peak of the target in the non-coherent accumulation result of these M sub-RD maps reaches the maximum. Perform threshold detection on the non-coherently accumulated sub-RD maps obtained in S5 respectively to obtain the range index I range and the Doppler index I dop_sub , and according to the range index I rangeDetermine the target distance information, and there is no limitation on using constant false alarm rate (CFAR) detection or other detection methods. It should be noted that the positions Int1, Int2, …, Int of the same target in each sub-RD diagram after non-coherent integration are Q the same.

[0048] Step S7: Compare the amplitudes of each target obtained in S5 in the Q sub-RD diagrams Int1, Int2, …, Int Q after non-coherent integration, and select the target with the largest amplitude as the real target; determine the real velocity of the target according to the serial number of the sub-RD diagram after non-coherent integration where the obtained real target is located and its Doppler index I dop_sub as follows:

[0049] Assume that the amplitude of the target in the sub-RD diagram Int q (q is an integer between 1 and Q) is the largest, then the Doppler dimension index corresponding to the real velocity of the target can be determined as I dop_total = I dop_sub +(q - 1)·N fft_dop / Q, where N fft_dop is the number of FFT points in the Doppler dimension; further determine the real velocity of the target according to I dop_total .

[0050] Step S8: Perform angle estimation on the target, which specifically includes the following steps:

[0051] Step S81: Combine the phase modulation methods of the transmission signals of different transmitting antennas TX during signal transmission, and determine the Doppler indices corresponding to the transmitting antennas TX1 to TXM according to the real Doppler index of the target.

[0052] Exemplarily, since the radar transmission signal to be processed in the present invention adopts a non-equidistant modulation DDMA-MIMO waveform, assume that at the transmitting end, the step size of the phase modulation of the transmitting antenna TXm between different PRTs is 2πb m / Q, and the Doppler index I dop_m corresponding to the transmitting antenna TXm can be expressed as: where, N fft_dop is the number of FFT points in the Doppler dimension, m = 1, 2, ..., M, and M is the number of transmitting antennas TX;

[0053] Step S82: According to the target distance index I rangeThe Doppler indices corresponding to different transmitting antennas TX, extract the data corresponding to different transmitting antennas TX in the RD maps corresponding to each receiving antenna RX, arrange the data values corresponding to the transceiver channels formed by different receiving antennas RX and transmitting antennas TX in sequence according to the antenna layout, obtain the signal of the virtual receiving antenna with a length of N·M, perform an FFT in the angle domain, and perform target detection in the angle dimension to obtain the target angle information, which will not be elaborated here.

[0054] Among them, in the steps S2 and S5, the non-coherent accumulation methods include but are not limited to taking the amplitude average, taking the modulus square average, taking the logarithm average, etc.

Claims

1. A radar signal processing method based on non-equispaced modulation DDMA-MIMO, characterized in that It includes the following steps: S1: Perform two-dimensional Fourier transform on the echo signals of different receiving channels to obtain N range-Doppler (RD) maps corresponding to N receiving channels respectively; N is the number of receiving channels; S2: Perform the first non-coherent integration on the N RD maps corresponding to N different receiving channels respectively to obtain the RD map after the first non-coherent integration; S3: According to the modulation method of the non-equispaced modulation DDMA-MIMO signal at the transmitting end, evenly divide the RD map after the first non-coherent integration into Q parts in the Doppler dimension to obtain Q sub-RD maps; where Q is the number of equal divisions of the Doppler dimension during DDMA modulation of the transmitted signal; S4: Extract M continuously distributed sub-RD maps from the Q sub-RD maps to form a sub-RD map combination, and a total of Q different sub-RD map combinations are formed; S5: For each sub-RD map combination, perform the second non-coherent integration on all M sub-RD maps to obtain the sub-RD map after non-coherent integration corresponding to each combination, and a total of Q sub-RD maps after non-coherent integration are obtained; S6: Perform threshold detection on the sub-RD maps after non-coherent integration obtained in S5 respectively to obtain the range index Irange and Doppler index Idop_sub of the target in each sub-RD map after non-coherent integration, and determine the target range information according to the range index Irange; S7: Compare the amplitudes of each target obtained in S5 in the Q sub-RD maps after non-coherent integration, select the target with the largest amplitude as the real target; determine the real speed of the target according to the serial number of the sub-RD map after non-coherent integration where the obtained real target is located and its Doppler index Idop_sub; S8: Estimate the angle of the target.

2. The radar signal processing method based on non-equispaced modulation DDMA-MIMO according to claim 1, wherein In step S4, the positions Int1, Int2, …, IntQ of the same target in each sub-RD map after non-coherent integration are the same.

3. A radar signal processing method based on non-equispaced modulation DDMA-MIMO according to claim 1, characterized in that, In the said step S7, the method for determining the real speed of the target is: Assume that the target has the maximum amplitude in the sub-RD diagram Int q Then the Doppler dimension index corresponding to the true velocity of the target is Idop_total = Idop_sub+(q - 1)·Nfft_dop / Q, where q is an integer between 1 and Q; Nfft_dop is the number of FFT points in the Doppler dimension; the true velocity of the target is determined according to Idop_total.

4. A radar signal processing method based on non-equidistant modulation DDMA-MIMO according to claim 1, characterized in that, In the said step S8, the method for estimating the angle of the target is: Step S81: Combine the phase modulation method of the transmitted signals of different transmitting antennas TX during signal transmission, and determine the Doppler indices corresponding to each transmitting antenna TX1~TXM according to the real Doppler index Idop_total of the target; Step S82: Index according to the target distance I range and the Doppler indices corresponding to different transmitting antennas TX, extract the data corresponding to different transmitting antennas TX in the RD map corresponding to each receiving antenna RX, and arrange the data values corresponding to the transceiver channels formed by different receiving antennas RX and transmitting antennas TX in sequence according to the antenna layout to obtain the signal of the virtual receiving antenna with a length of N·M, perform an angular domain Fourier transform on it, and perform target detection in the angle to obtain the target angle information.

5. A radar signal processing method based on non-equispaced modulation DDMA-MIMO as claimed in claim 4, characterized in that, In the said step S8, the method for determining the Doppler indices corresponding to each transmitting antenna TX1~TXM is: Assume that at the transmitting end, the step size of the modulation phase of the transmitting antenna TXm between different PRTs is 2πb m / Q, and the Doppler index I corresponding to the transmitting antenna TXm dop_m is expressed as: where N fft_dop is the number of FFT points in the Doppler dimension, m = 1, 2,..., M, and M is the number of transmitting antennas TX; b m takes an integer between 0 and M - 1, and b m ≤Q is determined according to the modulation phase.

6. A radar signal processing method based on non-equispaced modulation DDMA-MIMO according to claim 1, characterized in that, In the said steps S2 and S5, the non-coherent integration method adopts taking the average of amplitudes, taking the average of modulus squares, or taking the average of logarithms.

7. A radar signal processing method based on non-equispaced modulation DDMA-MIMO as described in claim 1, characterized in that, In the said step S6, the method for threshold detection is the constant false alarm rate (CFAR) detection method.

Images