Sublobe cancellation method based on range-doppler plane feature subspace clutter suppression

Abstract

The present application relates to a sidelobe cancellation method based on range-doppler plane feature subspace clutter suppression, belongs to the technical field of radar signal processing, and is particularly suitable for the field of radar with sidelobe interference suppression demand under the condition of clutter; the scene where the strong amplitude of clutter affects the sidelobe cancellation ratio index decline in the selection of interference samples in sidelobe cancellation. First, the radar intermediate frequency echo data is down-converted to baseband, arranged into a radar baseband complex pulse compression echo data matrix after pulse compression, then a threshold is set according to the modulus value of the noise data after baseband pulse compression, and an optimization problem function is further constructed; then the data is initialized and the low-rank iterative matrix and the sparse iterative matrix are solved according to the constraint threshold respectively; finally, the traditional radar sidelobe cancellation processing is carried out according to the final sparse iterative matrix after iteration; so as to effectively improve the radar sidelobe cancellation interference cancellation ratio, suppress the loss of target signal-to-noise ratio, and further significantly improve the target detection probability.

Description

Technical Field

[0001] This invention belongs to the field of radar signal processing technology. Background Technology

[0002] With the increasing maturity of modern phased array radar systems, using main and auxiliary channels for sidelobe cancellation to suppress or deceive interference entering from the sidelobes has become a common radar anti-jamming method. However, with the intelligent development of jammers and the influence of various strong clutter, the selection of interference samples has become a major problem. Inaccurate sample selection will further affect various key indicators such as the sidelobe cancellation ratio. To solve the problem of the sharp deterioration of the sidelobe cancellation interference suppression effect, it is necessary to adopt relevant measures to discard clutter samples and target samples, accurately identify interference samples, and calculate the sample correlation matrix of conventional main and auxiliary channels; thereby suppressing interference entering from the sidelobes, ensuring the target detection probability, and reducing the false alarm probability introduced by interference and clutter signals. Therefore, for the problem of radar sidelobe interference suppression under strong clutter background conditions, correctly selecting interference samples is a key technology for using sidelobe cancellation to suppress sidelobe interference and improve the radar target detection probability.

[0003] Wuhan Binhu Electronics Co., Ltd. disclosed a sidelobe cancellation method for radar in a strong clutter region in its invention patent application, "A Method for Sidelobe Cancellation in a Radar Strong Clutter Region" (Publication No.: CN114280552A, Application No.: CN202100545506.7). This method primarily suppresses clutter by performing MTD (Moving Target Detection) processing on the main and auxiliary channels separately, and then suppresses clutter again by performing traditional sidelobe cancellation on each channel. However, this method relies on Doppler characteristics for clutter suppression using moving target detection. If the target Doppler channel and the clutter Doppler channel overlap to some extent, the method will fail significantly. Furthermore, if the interference signal has Doppler characteristics, the interference sidelobes will enter the target Doppler channel region, severely affecting the target detection probability. Therefore, this method lacks universality.

[0004] The paper "Research on Sidelobe Cancellation of Sparse Array Radar Based on Measured Data" (Journal of Hefei University of Technology, 2012, Vol.35, pp:66-69) proposes using a "digital open-loop adaptive sidelobe cancellation system for weight calculation." However, this algorithm does not provide a detailed explanation of the specific sample selection, and it is also susceptible to the degradation of interference suppression performance under clutter conditions. Similarly, inevitably, for slow targets and clutter with large Doppler passbands, the lack of explicit constraints on the selection of interference samples leads to a sharp decline in interference suppression performance, thus affecting target detection. Summary of the Invention

[0005] To overcome the problems of difficulty in selecting interference samples, severe target signal-to-noise ratio loss, insufficient interference suppression capability, and complex engineering implementation in existing radar signal processing technologies, this invention proposes a sidelobe cancellation method based on clutter suppression in the range Doppler plane characteristic subspace. This invention does not disrupt the original signal processing flow, clearly distinguishes the clutter interference subspace, reduces target signal-to-noise ratio loss, and effectively suppresses sidelobe interference.

[0006] To achieve the above-mentioned technical objectives, the technical solution of the present invention includes:

[0007] First, the intermediate frequency radar echo data from the main and auxiliary channels of the radar receiver is acquired and down-converted to baseband. After baseband pulse compression, the data is arranged into a complex pulse compression echo data matrix A=[A1;A2;A3;…A2;A3;…A4] according to a coherent pulse group. M-1 A M ]; where matrix A is defined as an M-row N-column matrix, and each row vector A in the matrix is... i This represents the complex echo data of all range cells of a radar pulse, where i=1…M indicates that a radar echo coherent pulse group contains M pulses; the complex vector A i =[a1,a2,a3,…,a N-1 ,a N ] in i This represents the complex echo data corresponding to the i-th range cell in a radar pulse, where i=1…N indicates that a radar pulse contains a total of N range cells. The noise data magnitude after compression of the acquired radar baseband data pulse is set to... ε Therefore, the constraint threshold can be set as follows:

[0008] ;

[0009] Where parameters K It's a hyperparameter. P f This represents the desired false alarm probability for radar detection; construct the target optimization matrix. X = J + N + C + T Where C, T, J, and N represent the radar baseband complex pulse compressed echo data matrix containing clutter, target, interference, and noise components, respectively. Based on the sparsity of radar echo data and the strong correlation between clutter data across different range cells and pulses, the following constrained optimization problem function is further solved for the echo data matrix:

[0010] ;

[0011] Among the symbols This indicates the calculation of the nuclear norm of a matrix. This indicates finding the l-norm of a matrix, with parameters... λ This represents the regularization parameter; the constrained optimization problem described above can be further transformed into an unconstrained optimization problem function with a regularization parameter:

[0012] ;

[0013] Where parameters α as well as β Represents the regularization parameter, symbol This indicates calculating the 2-norm of a matrix, followed by initializing the collected data matrix. Y 0 represents the radar baseband complex pulse compressed echo data matrix A obtained after acquisition and processing, and is used to initialize the sparse iteration matrix. X 0=0; and solve the low-rank iteration matrix according to the following formula:

[0014] ;

[0015] in SVD μ (●) indicates a threshold value. μ The singular value decomposition, i.e., the corresponding singular values ​​are greater than or equal to μ Then stop the decomposition iteration, where k =1,2,3… D This represents the number of iterations; combining the low-rank iteration matrix obtained from the above steps, and solving for the sparse iteration matrix using the following formula. X k :

[0016] ;

[0017] in THR σ (●) indicates a threshold value. σ of l 2-norm decomposition, i.e., the corresponding l 2-norm greater than or equal to σ Then stop the decomposition iteration and set a threshold. σ This is a hyperparameter. The above low-rank and sparse iterative matrices first satisfy the iteration threshold and then the calculation stops; otherwise, iteration stops after the number of iterations reaches D. After the iteration stops, the low-rank iterative matrix can be obtained. Z k and sparse iterative matrix X k This allows us to target the sparse iteration matrix after the final iteration. X k The signal processing follows the traditional main and auxiliary channel radar sidelobe cancellation method, processing each pulse individually.

[0018] This invention can effectively improve the radar sidelobe interference cancellation ratio, suppress target signal-to-noise ratio loss, and further significantly improve the target detection probability, solving the target detection problem under conditions where clutter and interference exist. Attached Figure Description

[0019] Figure 1 This is a flowchart of the process of this invention.

[0020] Figure 2 This is the original time-frequency diagram containing clutter, noise, and the target in a specific embodiment of the present invention.

[0021] Figure 3 This is a time-frequency diagram processed by a sidelobe cancellation method based on range Doppler plane feature subspace clutter suppression in a specific embodiment of the present invention.

[0022] Figure 4 This is a comparison chart of the interference cancellation ratio between the sidelobe cancellation method based on range Doppler plane characteristic subspace clutter suppression in a specific embodiment of the present invention and the traditional sidelobe cancellation method. Detailed Implementation

[0023] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments, but this is not intended to limit the invention. A schematic diagram of the sidelobe cancellation method based on distance Doppler plane eigenspace clutter suppression of the present invention is shown below. Figure 1 As shown, the preferred implementation process is as follows:

[0024] Step 1: First, acquire the intermediate frequency radar echo data from the main and auxiliary channels of the radar receiver, and perform down-conversion processing to baseband; after baseband pulse compression, arrange the pulses into a complex pulse compression echo data matrix A=[A1;A2;A3;…A…]. M-1; A M ]; where matrix A is defined as an M-row N-column matrix, and each row vector A in the matrix is... i This represents the complex echo data of all range cells of a radar pulse, where i=1…M indicates that a radar echo coherent pulse group contains M pulses; the complex vector A i =[a1,a2,a3,…,a N-1 ,a N ] in i This represents the complex echo data corresponding to the i-th range cell in a radar pulse, where i=1…N indicates that a radar pulse contains a total of N range cells.

[0025] Step 2: Set the noise data modulus of the collected radar baseband data pulse compression to a value that is not specified. ε Therefore, the constraint threshold can be set as follows:

[0026] ;

[0027] Where parameters K It's a hyperparameter. P f This represents the desired false alarm probability for radar detection; construct the target optimization matrix. X = J + N + C + T Where C, T, J, and N represent the radar baseband complex pulse compressed echo data matrix containing clutter, target, interference, and noise components, respectively. Based on the sparsity of radar echo data and the strong correlation between clutter data across different range cells and pulses, the following constrained optimization problem function is further solved for the echo data matrix:

[0028] ;

[0029] Among the symbols This indicates the calculation of the nuclear norm of a matrix. This indicates finding the l-norm of a matrix, with parameters... λ This represents the regularization parameter; the constrained optimization problem described above can be further transformed into an unconstrained optimization problem function with a regularization parameter:

[0030] ;

[0031] Where parameters α as well as β Represents the regularization parameter, symbol This indicates the calculation of the 2-norm of a matrix.

[0032] Step 3: Initialize the data acquisition matrix Y 0 represents the radar baseband complex pulse compressed echo data matrix A obtained after acquisition and processing, and is used to initialize the sparse iteration matrix. X 0=0; and solve the low-rank iteration matrix according to the following formula:

[0033] ;

[0034] in SVD μ (●) indicates a threshold value. μ The singular value decomposition, i.e., the corresponding singular values ​​are greater than or equal to μ Then stop the decomposition iteration, where k =1,2,3… D Indicates the number of iterations.

[0035] Step 4: Using the low-rank iterative matrix obtained from the above steps, solve for the sparse iterative matrix using the following formula. X k:

[0036] ;

[0037] in THR σ (●) indicates a threshold value. σ of l 2-norm decomposition, i.e., the corresponding l 2-norm greater than or equal to σ Then stop the decomposition iteration and set a threshold. σ This is a hyperparameter.

[0038] Step 5: The calculation stops once the low-rank and sparse iterative matrices meet the iteration threshold; otherwise, iteration stops after the number of iterations reaches D. After iteration stops, the low-rank iterative matrix can be obtained. Z k and sparse iterative matrix X k This allows us to target the sparse iteration matrix after the final iteration. X k The signal processing follows the traditional main and auxiliary channel radar sidelobe cancellation method, processing each pulse individually.

[0039] The feasibility of the method of the present invention will be further verified through experimental simulation below.

[0040] In the experiment, it is assumed that the radar uses a narrowband linear frequency modulated (LFM) signal for target detection, with one main channel and one auxiliary channel for sidelobe cancellation. The simulation parameters are as follows: Assume that there are three point targets in the simulation scenario, namely target 1 (50 range gate, 5.5 velocity gate, 0dB scattering intensity), target 2 (150 range gate, 11.5 velocity gate, -20dB scattering intensity), and target 3 (100 range gate, 8.5 velocity gate, -10dB scattering intensity). The jamming released by the jammer is noise jamming. The jamming signal energy ratio (JSR) relative to the total energy of the three point targets is set to 20dB, and clutter is randomly generated to fill the entire pulse group. The signal-to-clutter ratio (SCR) is equal to 0dB. The average signal-to-noise ratio (SNR) of the signal is set to 5dB. Assuming a coherent pulse group has 128 pulses, each pulse has 400 effective range units, a signal bandwidth B of 10MHz, a pulse width of 10µs, and a baseband sampling rate of 20MHz, the simulation uses 128 sample points to calculate the weights for sidelobe cancellation. Based on the power calculations before and after interference suppression, the cancellation ratio of conventional sidelobe cancellation metrics is compared with the interference cancellation ratio improvement achieved using the method of this invention, ultimately analyzing the effectiveness of the proposed method.

[0041] Figure 2The image shows a scenario where clutter and interference are present in the original radar baseband data. As can be seen from the time-frequency diagram, the simulated three target signals are basically submerged by interference and clutter signals, and the clutter and interference signals are fused together, making it impossible to effectively distinguish between the clutter and interference signal components.

[0042] Figure 3 The image shows the time-frequency plot of radar baseband echo data after processing by the sidelobe cancellation method based on range Doppler plane eigenspace clutter suppression according to this invention. As can be seen from the figure, the signal-to-clutter ratio of the target is significantly improved after clutter preprocessing. The entire time-frequency plot is relatively smooth, and the clutter signal is effectively suppressed.

[0043] Figure 4 The figure shows a comparison of the interference cancellation ratios of radar baseband echo data processed by the sidelobe cancellation method based on range Doppler plane characteristic subspace clutter suppression according to this invention and those processed by conventional sidelobe cancellation. As can be seen from the figure, with the increase of clutter power, once the parameter clutter-to-noise ratio reaches above 15dB, the interference cancellation ratio of conventional sidelobe cancellation decreases significantly. However, the interference cancellation ratio of the method of this invention can still be maintained above 10dB, which fully verifies the effectiveness and feasibility of the method of this invention.

Claims

1. A sidelobe cancellation method based on range Doppler plane characteristic subspace clutter suppression, characterized in that: Step 1: Acquire intermediate frequency radar echo data from the main and auxiliary channels of the radar receiver, and perform down-conversion processing to baseband; After baseband pulse compression, the echo data matrix A = [A1;A2;A3;…A4] is formed by arranging the pulses into a coherent pulse group. M-1 A M ]; where matrix A is defined as an M-row N-column matrix, and each row vector A in the matrix is... i This represents the complex echo data of all range cells of a radar pulse, where i=1…M indicates that a radar echo coherent pulse group contains M pulses; the complex vector A i =[a1,a2,a3,…,a N-1 ,a N a in ] j This represents the complex echo data corresponding to the j-th range cell in a radar pulse, where j=1…N indicates that a radar pulse contains a total of N range cells; Step 2: Set the magnitude of the noise data after pulse compression of the acquired radar baseband data to be [value missing]. ε Set the constraint threshold as follows: ； Where parameters K It's a hyperparameter. P f Indicates the probability of a false alarm; Construct the objective optimization matrix X = J + N + C + T Where C, T, J, and N represent the radar baseband complex pulse compressed echo data matrix containing clutter, target, interference, and noise components, respectively. Based on the sparsity of radar echo data and the correlation between clutter data across different range cells and pulses, the constrained optimization problem function for solving the echo data matrix is ​​as follows: ； Among the symbols This indicates the calculation of the nuclear norm of a matrix. This indicates finding the l-norm of a matrix, with parameters... λ The regularization parameter represents the constraint optimization problem function, which is then transformed into an unconstrained optimization problem function with a regularization parameter. ； Where parameters α as well as β Represents the regularization parameter, symbol This indicates finding the 2-norm of a matrix; Step 3: Initialize the data acquisition matrix Y 0 represents the radar baseband complex pulse compressed echo data matrix A obtained after acquisition and processing, and is used to initialize the sparse iteration matrix. X 0=0, and solve for the low-rank iteration matrix according to the following formula: ； in SVD μ (●) indicates a threshold value. μ The singular value decomposition, i.e., the corresponding singular values ​​are greater than or equal to μ Then stop the decomposition iteration, where k =1,2,3… D Indicates the number of iterations; Step 4: Combine the above steps to solve for the low-rank iteration matrix and the sparse iteration matrix. X k : ； in THR σ (●) indicates a threshold value. σ of l 2-norm decomposition, i.e., the corresponding l 2-norm greater than or equal to σ Then stop the decomposition iteration and set a threshold. σ For hyperparameters; Step 5: In steps 3 and 4, first determine if the iteration threshold is met. If it is, stop the calculation; otherwise, determine if the iteration count has reached D. If it is, stop the iteration. After the iteration stops, the low-rank iteration matrix can be obtained. Z k and sparse iterative matrix X k ， This allows us to target the sparse iteration matrix after the final iteration. X k The signal processing follows the traditional main and auxiliary channel radar sidelobe cancellation method, processing each pulse individually.

Images