Radar target sidelobe suppression method based on cascade processing of APC and IARFT

Abstract

The present invention relates to a radar target sidelobe suppression method based on APC and IARFT cascade processing, which belongs to the field of radar signal processing, and particularly relates to radar target range sidelobe and velocity sidelobe suppression technology. The present invention first uses the RMVDR algorithm to perform adaptive pulse compression on each row of data in the pulse-range two-dimensional echo data of the coherent pulse train received by the radar, and then obtains the pulse-range two-dimensional output result after adaptive pulse compression; then, a velocity-range two-dimensional search grid is established, and the coherent pulse train data corresponding to each grid point is searched and extracted in the pulse-range two-dimensional output data matrix after adaptive pulse compression; finally, on the basis of iteratively adaptively solving the coherent accumulation weight vector corresponding to each grid point in the velocity-range two-dimensional search grid, combined with the coherent pulse train data corresponding to each grid point, the amplitude value estimation based on adaptive pulse compression and iterative adaptive RFT cascade processing corresponding to each grid point is calculated. Compared with the cascade processing mode of "conventional matched filtering" + "RFT", the radar target sidelobe suppression method based on adaptive pulse compression and iterative adaptive RFT cascade processing described in the present invention can suppress the problem of high range-velocity sidelobes generated by strong targets, avoid the obstruction of adjacent weak targets, and has the advantage of being easy to implement in engineering.

Description

Technical Field

[0001] The invention belongs to the field of radar signal processing, and in particular relates to radar target range sidelobe and speed sidelobe suppression technology. Background Art

[0002] When radar detects high-speed and weak targets such as stealth aircraft and cruise missiles, in order to improve the signal-to-noise ratio of such targets, a long-term coherent integration method is usually required. However, the long observation time, coupled with factors such as high target motion speed and high radar range resolution, will cause the target to move across range units. Traditional coherent integration methods, such as Moving Target Detection (MTD), implicitly assume that the target is only in one range unit during the coherent integration time. Therefore, when using the traditional MTD method for long-term coherent integration, it will lead to serious performance degradation. The long-term coherent integration method based on Radon Fourier Transform (RFT) can complete the coherent integration while compensating for the target's movement across range units according to the target's motion parameters, which better solves the problem of coherent accumulation and energy focusing of targets moving across range units.

[0003] However, the cascade processing of "conventional matched filtering" + "RFT" will cause strong targets to produce higher range-velocity sidelobes. These range-velocity sidelobes will block other nearby weak targets, thereby affecting the processing effects of subsequent target detection and tracking. In particular, when the conventional matched filtering processing does not suppress the range sidelobes of strong targets well, the velocity focusing performance of the RFT method will be further deteriorated.

[0004] In order to solve the above problems, the present invention provides a radar target sidelobe suppression method based on cascade processing of adaptive pulse compression (Adaptive Pulse Compression, APC) and iterative adaptive RFT (Iterative Adaptive Radon Fourier Transform, IARFT). Summary of the invention

[0005] The purpose of the present invention is to provide a radar target sidelobe suppression method based on APC and IARFT cascade processing, aiming at the problem that the strong target range-velocity sidelobe blocks other adjacent weak targets caused by matched filtering and RFT cascade processing in the traditional radar coherent processing process, wherein the technical problems to be solved include:

[0006] (1) Suppression of strong target range sidelobes;

[0007] (2) The problem of suppressing strong target velocity sidelobes.

[0008] The radar target sidelobe suppression method based on APC and IARFT cascade processing of the present invention is characterized by comprising the following technical measures:

[0009] Step 1: First, the pulse-range two-dimensional echo data of the radar coherent pulse train is received, and then the RMVDR method is used to perform adaptive pulse compression processing on the range-dimensional echo of each pulse to solve the problem of strong target range sidelobe suppression;

[0010] Step 2: for each grid point in the speed-distance two-dimensional search grid, search and extract the coherent pulse train data corresponding to each grid point in the pulse-distance two-dimensional data matrix after the adaptive pulse compression process;

[0011] Step 3: For each grid point in the speed-distance two-dimensional search grid, use the coherent pulse train data corresponding to the grid point obtained in step 2 to calculate the coherent accumulation weight vector required for iterative adaptive RFT processing, and then obtain the coherent accumulation result corresponding to the grid point; since iterative adaptive RFT processing is used, this processing process suppresses the velocity sidelobes during coherent accumulation.

[0012] Compared with the cascade processing mode of "conventional matched filtering" + "RFT", the radar target sidelobe suppression method based on the cascade processing of APC and IARFT described in the present invention has the following beneficial effects:

[0013] (1) This method solves the problem that strong targets are prone to generate high range-velocity sidelobes in the cascade processing of "conventional matched filtering" + "RFT", avoiding the occlusion of adjacent weak targets.

[0014] (2) This method adopts the cascade processing method of "adaptive pulse compression" + "iterative adaptive RFT", which is convenient for engineering implementation while suppressing the range-velocity sidelobes of strong targets and obtaining high range-velocity focusing capability. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a flow chart of a radar target sidelobe suppression method based on APC and IARFT cascade processing.

[0016] Figure 2 It is a schematic diagram used to explain the meaning of indicator vector. DETAILED DESCRIPTION

[0017] The present invention will be described in further detail below in conjunction with the accompanying drawings. Figure 1 The specific implementation of the present invention is divided into the following steps:

[0018] (1) Adaptive pulse compression is performed on each row of data in the pulse-range two-dimensional echo data of the coherent pulse train received by the radar using the RMVDR algorithm, thereby obtaining a pulse-range two-dimensional output result after adaptive pulse compression;

[0019] (2) establishing a velocity-distance two-dimensional search grid, searching and extracting the coherent pulse train data corresponding to each grid point in the pulse-distance two-dimensional output data matrix after adaptive pulse compression;

[0020] (3) Based on the iterative and adaptive solution of the coherent accumulation weight vector corresponding to each grid point in the speed-distance two-dimensional search grid, the amplitude value estimation corresponding to each grid point based on adaptive pulse compression and iterative adaptive RFT cascade processing is calculated and combined with the coherent pulse train data corresponding to each grid point.

[0021] The following is a detailed description of the above steps:

[0022] (1) The pulse-range two-dimensional echo data of the coherent pulse train received by the radar is recorded as X. Then, for each row in X, that is, the distance dimension echo data corresponding to each pulse, the RMVDR algorithm is used to perform adaptive pulse compression, thereby obtaining the pulse-range two-dimensional output result after adaptive pulse compression, which is recorded as The specific operations of this step are as follows:

[0023] ① The pulse-range two-dimensional echo data X of the coherent pulse train received by the radar is an M×(D+7(N-1))-dimensional data matrix, where M is the number of coherent pulses, (D+7(N-1)) is the number of distance sampling points, the distance sampling interval between two adjacent points is equal to the distance resolution △r of the transmitted waveform, D is the number of sampling points in the range of interest, and N is the number of sampling points in the pulse; the (m,d)th element in the data matrix X is recorded as x m,d , then, the matrix X is expressed as follows, where 1≤m≤M, 1≤d≤(D+7(N-1)),

[0024]

[0025] ② For the echo data x of the mth (1≤m≤M) pulse m,1-3(N-1):D+4(N-1) , use the transmitted signal s to perform matched filtering, and the output result is recorded as

[0026]

[0027] Wherein, the superscript (0) indicates the 0th iteration, and the transmitted signal s is expressed as, s = [s(0), s(1), ..., s(N-1)] T , the superscript T indicates the transpose of a vector;

[0028] ③ Enter the lth iteration of adaptive pulse compression (1≤l≤3), for the dth distance sampling point, 1-(3-l)(N-1)≤d≤D+(3-l)(N-1), use x m,d:d+N-1 =[x m,d x m,d+1 …x m,d+N-1 ] represents the vector consisting of N distance sampling points starting from the dth distance sampling point; then for the vector x m,d:d+N-1 The weight vector required to implement the lth adaptive pulse compression for,

[0029]

[0030] Wherein, the superscript (l) indicates the lth iteration, and the subscript (m, d) indicates the targeted grid point; represents the echo covariance matrix, is the noise power level and can be measured by the radar system, I N is an N×N dimensional unit matrix; is the signal covariance matrix within the pulse, in, It represents the amplitude value estimate of the d+nth distance sampling point obtained at the mth pulse and the l-1th iteration. When l = 1, that is s n represents an N×1-dimensional shift vector based on s,

[0031]

[0032] ④Calculation The amplitude estimate of the mth pulse and the dth distance sampling point at the lth iteration can be obtained. Where, the superscript H represents the conjugate transpose of the vector;

[0033] ⑤Let d=d+1, repeat steps ③ to ④ until d=D+(3-1)(N-1); get

[0034] ⑥ Let l = l + 1, repeat steps ③ to ⑤ until l = 3 and stop; the adaptive pulse compression processing result corresponding to the mth pulse is obtained, which is recorded as

[0035] ⑦ Let m = m + 1, repeat steps ② to ⑥ until m = M; get And use this to construct the data matrix after adaptive pulse compression processing, recorded as and used as input data for iterative adaptive RFT processing,

[0036]

[0037] (2) Establishing a speed-distance two-dimensional search grid, searching and extracting the coherent pulse train data corresponding to each grid point in the pulse-distance two-dimensional output data matrix after adaptive pulse compression; the specific operation of this step is as follows:

[0038] ① Establish a distance search grid

[0039] Using the range resolution △r of the transmitted waveform and the number of range sampling points D of the echo data, a range search grid is established, where the distance value corresponding to the dth range grid point is,

[0040] ②Establish a speed search grid

[0041] Use v max represents the maximum speed of the target of interest; the size of the speed grid is denoted as Δv, Where λ represents the carrier wavelength, and T represents the pulse repetition period. Thus, a velocity search grid is established, where the velocity value corresponding to the qth velocity grid point is, <·> indicates rounding operation;

[0042] ③ Using the distance search grid points formed by ① and the speed search grid points formed by ②, a speed-distance two-dimensional search grid is formed;

[0043] ④ According to each speed-distance two-dimensional search grid (q, d), 1≤q≤Q, 1≤d≤D, according to the subscripts of each element in the following formula Search in the data matrix to obtain the coherent pulse train data corresponding to the grid point, denoted as z (q,d) , 1≤q≤Q, 1≤d≤D,

[0044]

[0045] in represents the rounding operation, and the subscript (q, d) represents the grid point targeted;

[0046] (3) On the basis of iteratively and adaptively solving the coherent accumulation weight vector corresponding to each grid point in the speed-distance two-dimensional search grid, combined with the coherent pulse train data corresponding to each grid point, the amplitude value estimation based on adaptive pulse compression and iterative adaptive RFT cascade processing corresponding to each grid point is calculated; the specific operation of this step is as follows:

[0047] ① Use the speed-distance two-dimensional search grid (q, d) to find the speed value v q Deterministic Doppler filter The coherent pulse train data z corresponding to the grid point(q,d) Perform coherent accumulation and obtain Represents the amplitude value estimation corresponding to the grid point (q, d) of the p = 0th iteration adaptive RFT processing output; traversing all the speed-distance two-dimensional search grid points, the amplitude value estimation matrix of the 0th iteration adaptive RFT processing output can be obtained

[0048]

[0049] ② When implementing the pth iteration of the adaptive RFT process, the coherent accumulation weight vector corresponding to the grid point (q, d) is first calculated. Then calculate The amplitude value estimate corresponding to the velocity-distance grid point (q, d) at the pth iteration can be obtained in, represents the echo covariance matrix between pulses, I M is the M×M dimensional unit matrix, represents the signal covariance matrix between pulses, in, represents the sum of all coherent pulse train data related to the velocity-distance grid (q, d) at the pth iteration,

[0050]

[0051] The superscript (p) indicates the pth iteration. represents the Hadamard product, represents the amplitude value estimate of the velocity-distance grid point (j, k) obtained at the p-1th iteration. When p = 1, that is indicator[(q,d),(j,k)] represents an M×1 dimensional indicator vector, which is used to describe the coherent pulse train data z corresponding to the grid point (j,k) (j,k) The coherent pulse train data z at the grid point (q, d) (q,d) overlap between them;

[0052] The following is based on the drawings in the specification Figure 2 , explains the meaning of the indicator vector: the coherent pulse train data corresponding to the speed-distance grid point (q, d) in the data matrix The location in the instruction manual is as follows Figure 2 As shown in the shaded area from the upper left to the lower right in the figure, the coherent pulse train data corresponding to the speed-distance grid point (j, k) is in the data matrix The location in the instruction manual is as follows Figure 2As shown in the shaded area from the upper right to the left; it can be seen from the figure that the shaded area from the upper left to the lower right overlaps with the shaded area from the upper right to the left at the third and fourth pulses. Therefore, the M×1-dimensional indicator vector indicator[(q,d),(j,k)] is obtained as follows:

[0053]

[0054] That is, the 3rd and 4th elements are 1, and the rest are 0;

[0055] ③ Let d = d + 1, repeat step ② until d = D; get the vector formed by the D amplitude value estimation corresponding to the p-th iteration RFT processing and the q-th velocity grid

[0056] ④ Let q = q + 1, repeat steps ② to ③ until q = Q; then we get That is, the amplitude value estimation matrix of the p-th iteration adaptive RFT processing output;

[0057] ⑤Let p=p+1, repeat steps ② to ④ until p=7; then we get This is the final processing result of the radar target sidelobe suppression method based on the cascade processing of APC and IARFT.

Claims

1. Radar target sidelobe suppression method based on APC and IARFT cascade processing, It is characterized in that The following steps are involved: S1. For each row of data in the pulse-range two-dimensional echo data of the coherent pulse train received by the radar, the RMVDR algorithm is used to perform adaptive pulse compression, thereby obtaining a pulse-range two-dimensional output result after adaptive pulse compression; S2, establishing a speed-distance two-dimensional search grid, searching and extracting the coherent pulse train data corresponding to each grid point in the pulse-distance two-dimensional output data matrix after adaptive pulse compression; S3, based on iteratively and adaptively solving the coherent accumulation weight vector corresponding to each grid point in the speed-distance two-dimensional search grid, combined with the coherent pulse train data corresponding to each grid point, calculate the amplitude value estimation based on adaptive pulse compression and iterative adaptive RFT cascade processing corresponding to each grid point, specifically including: S31, using the speed-distance two-dimensional search grid point (q, d) velocity value v q Deterministic Doppler filter The coherent pulse train data z corresponding to the grid point (q,d) Perform coherent accumulation and obtain Represents the amplitude value estimation corresponding to the grid point (q, d) of the p = 0th iteration adaptive RFT processing output; traversing all the speed-distance two-dimensional search grid points, the amplitude value estimation matrix of the 0th iteration adaptive RFT processing output can be obtained S32, when the adaptive RFT process is implemented for the pth iteration, the coherent accumulation weight vector corresponding to the grid point (q, d) is first calculated, Then calculate The amplitude value estimate corresponding to the velocity-distance grid point (q, d) at the pth iteration can be obtained in, represents the echo covariance matrix between pulses, I M is the M×M dimensional unit matrix, represents the signal covariance matrix between pulses, in, represents the sum of all coherent pulse train data related to the velocity-distance grid (q, d) at the pth iteration, The superscript (p) indicates the pth iteration. represents the Hadamard product, represents the amplitude value estimate of the velocity-distance grid point (j, k) obtained at the p-1th iteration. When p = 1, that is indicator[(q,d),(j,k)] represents an M×1 dimensional indicator vector, which is used to describe the coherent pulse train data z corresponding to the grid point (j,k) (j,k) The coherent pulse train data z at the grid point (q, d) (q,d) overlap between them; The coherent pulse train data corresponding to the speed-distance grid point (q, d) is in the data matrix The coherent pulse train data corresponding to the position and velocity-distance grid point (j, k) in the data matrix The positions in overlap at the 3rd and 4th pulses, so the M×1-dimensional indicator vector indicator[(q,d),(j,k)] is obtained as follows: That is, the 3rd and 4th elements are 1, and the rest are 0; S33, let d = d + 1, repeat step S32 until d = D, and then obtain the vector formed by the D amplitude values ​​estimated for the p-th iteration RFT processing and the q-th velocity grid. S34, let q = q + 1, repeat steps S32 to S33 until q = Q, then stop; get That is, the amplitude value estimation matrix of the p-th iteration adaptive RFT processing output; S35, let p=p+1, repeat steps S32 to S34 until p=7; get This is the final processing result of the radar target sidelobe suppression method based on the cascade processing of APC and IARFT.

2. The radar target sidelobe suppression method based on APC and IARFT cascade processing according to claim 1, It is characterized in that The step S1 specifically includes: S11. The pulse-range two-dimensional echo data X of the coherent pulse train received by the radar is an M×(D+7(N-1))-dimensional data matrix, where M is the number of coherent pulses, (D+7(N-1)) is the number of range sampling points, the range sampling interval between two adjacent points is equal to the range resolution △r of the transmitted waveform, D is the number of sampling points in the range segment of interest, and N is the number of sampling points in the pulse; The (m,d)th element in the data matrix X is denoted by x m,d , then, the matrix X is expressed as follows, where 1≤m≤M, 1≤d≤(D+7(N-1)), S12, echo data x for the mth (1≤m≤M) pulse m,1-3(N-1):D+4(N-1) , use the transmitted signal s to perform matched filtering, and the output result is recorded as Wherein, the superscript (0) indicates the 0th iteration, and the transmitted signal s is expressed as, s = [s(0), s(1), ..., s(N-1)] T , the superscript T indicates the transpose of a vector; S13, enter the lth iteration of adaptive pulse compression (1≤l≤3), for the dth distance sampling point, 1-(3-l)(N-1)≤d≤D+(3-l)(N-1), use x m,d:d+N-1 =[x m,d x m,d+1 …x m,d+N-1 ] represents the vector consisting of N distance sampling points starting from the dth distance sampling point; then for the vector x m,d:d+N-1 The weight vector required to implement the lth adaptive pulse compression for, Wherein, the superscript (l) indicates the lth iteration, and the subscript (m, d) indicates the targeted grid point; represents the echo covariance matrix, is the noise power level and can be measured by the radar system, I N is an N×N dimensional unit matrix; is the signal covariance matrix within the pulse, in, It represents the amplitude value estimate of the d+nth distance sampling point obtained at the mth pulse and the l-1th iteration. When l = 1, that is s n represents an N×1-dimensional shift vector based on s, S14. Calculation The amplitude estimate of the mth pulse and the dth distance sampling point at the lth iteration can be obtained. Where, the superscript H represents the conjugate transpose of the vector; S15, let d = d + 1, repeat steps S13 to S14 until d = D + (3-1) (N-1) and stop; get S16, let l=l+1, repeat steps S13 to S15 until l=3; obtain the adaptive pulse compression processing result corresponding to the mth pulse, recorded as S17, let m=m+1, repeat steps S12 to S16 until m=M; get And use this to construct the data matrix after adaptive pulse compression processing, recorded as and used as input data for iterative adaptive RFT processing, 3. The radar target sidelobe suppression method based on APC and IARFT cascade processing according to claim 1, It is characterized in that The step S2 specifically includes: S21, establish a distance search grid; use the range resolution △r of the transmitted waveform and the number of sampling points D in the range segment of interest to establish a distance search grid, where the distance value corresponding to the dth range grid point is, S22, establish a speed search grid; use v max represents the maximum speed of the target of interest; the size of the speed grid is denoted as Δv, Where λ represents the carrier wavelength, and T represents the pulse repetition period. Thus, a velocity search grid is established, where the velocity value corresponding to the qth velocity grid point is, <·> indicates rounding operation; S23, using the distance search grid points formed in S21 and the speed search grid points formed in S22 to form a speed-distance two-dimensional search grid; S24, according to each speed-distance two-dimensional search grid (q, d), 1≤q≤Q, 1≤d≤D, according to the subscripts of the elements in the following formula Search in the data matrix to obtain the coherent pulse train data corresponding to the grid point, denoted as z (q,d) , 1≤q≤Q, 1≤d≤D, in represents the rounding up operation, and the subscript (q, d) represents the grid point being targeted.

Images