A correlation-based method for suppressing sidelobes in frequency-agile radar

Abstract

This invention discloses a correlation-based sidelobe suppression method for frequency-agile radar. It constructs an echo signal model of a coarse-resolution range cell in a sparse scene of frequency-agile radar, including the observed signal and a dictionary matrix. The echo signal undergoes inter-pulse coherent processing to obtain velocity and range information. A sidelobe suppression matrix is ​​obtained using the correlation between the echo signal and the dictionary matrix. This sidelobe suppression matrix is ​​then used to suppress the sidelobes of the coherently processed signal, achieving target detection. This invention features simple steps, clear logic, and can effectively suppress sidelobes and accurately recover the target.

Description

Technical Field

[0001] This invention relates to the field of radar signal processing, and in particular to a correlation-based method for suppressing sidelobes in frequency-agile radar. Background Technology

[0002] In recent years, the electromagnetic environment in which modern radar operates has become increasingly complex, seriously threatening its battlefield survivability. Frequency-agile radar can change its signal form in real time according to the electromagnetic environment and target characteristics, thereby achieving better detection capabilities, low intercept performance, and anti-jamming performance. However, the random switching of the radar signal carrier frequency between pulses poses difficulties for coherent synthesis in signal processing. Applying compressed sensing algorithms to frequency-agile radar can reconstruct high-resolution range-velocity target scenes by constructing an observation matrix. However, real-world scenes contain a large amount of noise, and the reconstruction using compressed sensing algorithms often contains numerous sidelobes, which can mask weak targets or lead to false alarms, severely affecting the reconstruction performance of the target scene. Summary of the Invention

[0003] The purpose of this invention is to propose a correlation-based sidelobe suppression method for frequency agile radar, which can suppress target sidelobes in the coherent processing results of frequency agile radar, improve the detection probability of targets, and reduce the false alarm rate.

[0004] The technical solution to achieve the objective of this invention is as follows: a correlation-based method for suppressing sidelobes in frequency-agile radar, comprising the following steps:

[0005] Step 1: Construct an echo signal model for a coarse-resolution range cell in a sparse scenario of frequency-agile radar, including the observed signal and dictionary matrix;

[0006] Step 2: Perform inter-pulse coherent processing on the echo signal to obtain velocity and distance information;

[0007] Step 3: Obtain the sidelobe suppression matrix by utilizing the correlation between the echo signal and the dictionary matrix;

[0008] Step 4: Use a sidelobe suppression matrix to suppress the sidelobes of the coherently processed signal to achieve target detection.

[0009] Step 1: Construct an echo signal model for a coarse-resolution range cell in a sparse scene of frequency-agile radar, including the observed signal and dictionary matrix. The specific method is as follows:

[0010] (A1) If there are G targets in the coarse-resolution range cell of interest, then the echo signal S of the nth pulse in one CPI of the frequency-agile radar is... r (n) is:

[0011]

[0012] Where, β g Let r represent the scattering coefficient of the g-th target. g and v g T represents its distance and radial velocity, respectively. r f represents the pulse repetition period. c Indicates the initial carrier frequency, d n Δf represents the carrier frequency hopping codeword, Δf represents the frequency hopping interval, and c represents the speed of light.

[0013] (A2) The echo signal model is simplified to:

[0014]

[0015] in, γ g denoted by scattering coefficient, p represents range phase factor, and q represents velocity phase factor;

[0016] (A3) The echo signal S of N pulses within one CPI r (n), n∈{0,1,2,…N-1} constitute the observation signal:

[0017] y-[S(0),S(1),S(2),…,S(N-1)] T

[0018] (A4) The echo signal S obtained from (A2) r (n) Discretization:

[0019]

[0020] Where, γ k,l Indicates that it is located in the grid (p) k q l ) target scattering intensity;

[0021] make but

[0022]

[0023] (A5) Discretize S from (A4) r Substituting (n) into the observed signal in (A3) yields an echo signal within CPI:

[0024]

[0025] Simplified, we get:

[0026] y = Ax

[0027] Where y is the observed signal within one CPI, x = [γ] 0,0 γ 0,1 …γK-1,L-1 ] T Let A be the signal to be recovered, and let A be the dictionary matrix.

[0028]

[0029] Step 2 involves performing inter-pulse coherent processing on the echo signal to obtain velocity and range information. This involves reconstructing the signal x to be recovered using a compressed sensing algorithm. The non-zero elements in x contain the high-resolution range and velocity information of the target. The compressed sensing algorithm utilizes the amplitude and phase information of the echo. The process of reconstructing x is essentially solving a convex optimization problem. The specific process is as follows:

[0030] (B1) Construct a convex optimization problem based on the observed signal and dictionary matrix:

[0031] min|x|1, sty=Ax+ω

[0032] Where ω is noise, x is the scattering coefficient vector of the target in the observation space, and is the signal to be recovered;

[0033] (B2) Use the MATLAB Convex Optimization Toolbox CVX or the Orthogonal Matching Pursuit algorithm to solve the (B1) convex optimization problem to obtain the solution x. The non-zero elements in x contain the high-resolution distance and velocity information of the target.

[0034] Step 3: Obtain the sidelobe suppression matrix by utilizing the correlation between the echo signal and the dictionary matrix. The specific method is as follows:

[0035] (C1) Calculate the correlation b between the echo signal and the dictionary matrix:

[0036] b = A H ·y;

[0037] Where A is the dictionary matrix, y is the observed signal, and the correlation b is a column vector of length KL;

[0038] (C2) Based on the correlation, the sidelobe suppression matrix W is obtained:

[0039]

[0040] Where b0, b1, ... b KL-1 The element is in b.

[0041] Step 4: Use a sidelobe suppression matrix to suppress the sidelobes of the coherently processed signal to achieve target detection. The specific method is as follows:

[0042] (D1) Modulate the reconstructed signal x using the sidelobe suppression matrix, i.e., x' = W·x;

[0043] (D2) Use the maximum inter-class variance method to obtain the threshold T of x';

[0044] (D3) Set the values ​​of x' less than T to zero to obtain a one-dimensional high-resolution range-velocity vector x".

[0045] (D4) Convert the one-dimensional high-resolution range-velocity vector x" into a two-dimensional range-velocity matrix to obtain the target range and velocity information after sidelobe suppression.

[0046] A correlation-based frequency agile radar sidelobe suppression system is provided, which achieves correlation-based sidelobe suppression of frequency agile radar based on the aforementioned frequency agile radar sidelobe suppression method.

[0047] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements correlation-based sidelobe suppression for frequency-agile radar based on the aforementioned frequency-agile radar sidelobe suppression method.

[0048] A computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, correlation-based sidelobe suppression of frequency-agile radar is achieved based on the aforementioned frequency-agile radar sidelobe suppression method.

[0049] Compared with existing technologies, the significant advantages of this invention are: 1) It uses compressed sensing algorithms to reconstruct targets without needing to know the number of targets in the scene, making it more realistic. 2) To address the problem of excessive sidelobes in the reconstructed targets, it introduces the correlation between the observed signal and the dictionary matrix to construct a sidelobe suppression matrix, achieving sidelobe suppression with strong robustness and low computational complexity. Attached Figure Description

[0050] Figure 1 This is a flowchart of the frequency-agile radar sidelobe suppression method of the present invention.

[0051] Figure 2 This is a graph showing the results of coherent processing in the simulation experiment.

[0052] Figure 3 This is a diagram showing the results after sidelobe suppression in the simulation experiment. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0054] A correlation-based method for suppressing sidelobes in frequency-agile radar includes the following steps:

[0055] Step 1: Construct a coarse-resolution range cell with G targets in a sparse scene of the frequency agile radar. The echo signal of the nth pulse in one CPI of the frequency agile radar is:

[0056]

[0057] Where, β g Let r represent the scattering coefficient of the g-th target. g and v g T represents its distance and radial velocity, respectively. r f represents the pulse repetition period. c Indicates the initial carrier frequency, d n Δf represents the carrier frequency switching codeword, Δf represents the frequency switching interval, and c represents the speed of light.

[0058] Simplify the signal, Substituting the echo signal into the model of the observed signal, we obtain the following:

[0059]

[0060] Where, γ g denoted by scattering coefficient, p represents range phase factor, and q represents velocity phase factor;

[0061] The range and velocity elements of interest are discretized into K×L meshes.

[0062]

[0063] Where γ k,l Indicates that it is located in the grid (p) k q l The target scattering intensity at (p) k q l If there is a target at position ) then γ k,l If there is a numerical value, then γ k,l It is zero.

[0064] scattering intensity matrix Vectorization yields the scattering coefficient vector.

[0065] Define the guidance vector The echo signal can then be rewritten as:

[0066] S r (n)=[a0,a1,…,a KL-1 ]·x

[0067] For a coarse-resolution range cell, the echo signals of N pulses within one CPI constitute the observation signal:

[0068] y=[S(0),S(1),S(2),…,S(N-1)] T

[0069] The echo signal model within one CPI is:

[0070]

[0071] Abbreviated as:

[0072] y = Ax + ω

[0073] Where ω represents noise. It is a dictionary matrix, that is

[0074]

[0075] Step 2: Perform inter-pulse coherent processing on the echo signal to obtain velocity and distance information;

[0076] x is reconstructed using compressed sensing, where the non-zero elements of x contain high-resolution range and velocity information of the target. Compressed sensing fully utilizes the amplitude and phase information of the echo, therefore the method is coherent. The process of reconstructing x is essentially solving a convex optimization problem.

[0077] min|x|1, sty=Ax+ω

[0078] Compressed sensing reconstruction algorithms can be implemented using the MATLAB convex optimization toolbox CVX or the orthogonal matching pursuit algorithm.

[0079] In step (C), the correlation between the echo signal and the dictionary matrix is ​​first calculated, i.e., b = A. H ·y;

[0080] Then, the sidelobe suppression matrix is ​​obtained based on the correlation.

[0081] Where b0, b1, ... b KL-1 For elements in b;

[0082] Step (D) uses a sidelobe suppression matrix to suppress the sidelobes of the coherently processed signal to achieve target detection.

[0083] x'=W·x

[0084] The threshold T of the solution after sidelobe suppression matrix processing is obtained by using the Otsu' method. Values ​​in x' less than T are set to zero. Then, the one-dimensional high-resolution range-velocity vector x" is converted into a two-dimensional range-velocity matrix, thus obtaining the target range and velocity information after sidelobe suppression.

[0085] In summary, the method of the present invention has simple steps, clear logic, and can effectively suppress sidelobes and accurately recover the target.

[0086] Example

[0087] To verify the effectiveness of the present invention, the following experiment was conducted.

[0088] The frequency-agile radar transmits a rectangular signal with 64 pulses per CPI. The initial frequency of the radar frequency hopping is 9 GHz, the number of hopping points is 64, the hopping interval is 15 MHz, the synthetic bandwidth is 960 MHz, and the pulse repetition frequency is 25 kHz. Assume there are three moving point targets in the radar observation scene, with their high-resolution range-velocity corresponding to grid points (7, 34), (5, 27), and (22, 24), respectively, and all having a scattering coefficient of 1. The signal-to-noise ratio is -5 dB.

[0089] (A) The echo signal of the coarse-resolution range cell of interest is constructed based on the simulation parameters as follows:

[0090]

[0091] The observation signal is then formed by the echo signal of 64 pulses within a CPI.

[0092] y=[S(0),S(1),S(2),…,S(63)] T

[0093] Construct dictionary matrix

[0094] Step (B) uses the orthogonal matching pursuit algorithm as the recovery algorithm of the compressed sensing model to reconstruct the target.

[0095] Step (C) calculates the correlation between the echo signal and the dictionary matrix, and constructs the sidelobe suppression matrix.

[0096]

[0097] Step (D) uses a sidelobe suppression matrix to suppress the sidelobes of the coherently processed signal, x' = W·x. The threshold T is obtained using the graythresh function in MATLAB. Values ​​in x' less than T are set to zero. Then, the one-dimensional high-resolution range-velocity vector x" is converted into a range-velocity two-dimensional matrix, thus obtaining the target range and velocity information after sidelobe suppression.

[0098] Based on the above simulation experimental conditions and process, the following results were obtained: Figure 1-2 The comparison is as follows. Figure 1 This is the result of coherent processing under normal circumstances, from Figure 1 As can be seen, normal coherent processing can detect many targets, resulting in false alarms. Figure 2 This is the result of sidelobe matrix suppression in this invention, from... Figure 2As can be seen, only the three preset targets were accurately recovered, while false targets caused by other sidelobes were suppressed. This demonstrates that the present invention can effectively suppress target sidelobes in the coherent processing of frequency-agile radar.

[0099] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0100] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these modifications and improvements all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A correlation-based method for suppressing sidelobes in frequency-agile radar, characterized in that, The steps are as follows: Step 1: Construct an echo signal model for a coarse-resolution range cell in a sparse scenario of frequency-agile radar, including the observed signal and dictionary matrix; Step 2: Perform inter-pulse coherent processing on the echo signal to obtain velocity and distance information; Step 3: Obtain the sidelobe suppression matrix by utilizing the correlation between the echo signal and the dictionary matrix; Step 4: Use a sidelobe suppression matrix to suppress the sidelobes of the coherently processed signal to achieve target detection; in: Step 1: Construct an echo signal model for a coarse-resolution range cell in a sparse scene of frequency-agile radar, including the observed signal and dictionary matrix. The specific method is as follows: (A1) If there are G targets in the coarse-resolution range cell of interest, then the echo signal of the nth pulse in one CPI of the frequency-agile radar is... for: ； in, This represents the scattering coefficient of the g-th target. and These represent its distance and radial velocity, respectively. Indicates the pulse repetition period. Indicates the initial carrier frequency. Indicates a carrier frequency hopping codeword. Indicates the frequency transition interval. Represents the speed of light; (A2) The echo signal model is simplified to: ； in, , Represents the scattering coefficient. Represents the distance phase factor. Represents the velocity phase factor; (A3) Echo signal of N pulses within one CPI Constituting the observation signal: Y=[S r (0),S r (1),S r (2),…,S r (N-1)] T ; (A4) Discretize the range and velocity elements of interest into a K×L mesh: ； in, Indicates location within the grid Target scattering intensity at the location; make ,but ； (A5) Discretize the data in (A4) Substituting the observed signal into (A3) yields an echo signal within a CPI: ； Simplified, we get: ； in For an observed signal within one CPI, This is a signal that is yet to be recovered. For dictionary matrix: ； Step 2 involves performing inter-pulse coherent processing on the echo signal to obtain velocity and range information. This involves reconstructing the signal x to be recovered using a compressed sensing algorithm. The non-zero elements in x contain the high-resolution range and velocity information of the target. The compressed sensing algorithm utilizes the amplitude and phase information of the echo. The process of reconstructing x is essentially solving a convex optimization problem. The specific process is as follows: (B1) Construct a convex optimization problem based on the observed signal and dictionary matrix: ； in Let be noise, x be the scattering coefficient vector of the target in the observation space, and be the signal to be recovered; (B2) Use the MATLAB Convex Optimization Toolbox CVX or the Orthogonal Matching Pursuit algorithm to solve the (B1) convex optimization problem to obtain the solution x. The non-zero elements in x contain the high-resolution distance and velocity information of the target. Step 3: Obtain the sidelobe suppression matrix by utilizing the correlation between the echo signal and the dictionary matrix. The specific method is as follows: (C1) Calculate the correlation between the echo signal and the dictionary matrix. : ； in It is a dictionary matrix. For the observed signal, the correlation b is a column vector of length KL; (C2) Based on the correlation, the sidelobe suppression matrix W is obtained: ； in For elements in b; Step 4: Use a sidelobe suppression matrix to suppress the sidelobes of the coherently processed signal to achieve target detection. The specific method is as follows: (D1) Modulation of the reconstructed signal x is performed using the sidelobe suppression matrix, i.e. ; (D2) Obtain using the Otsu's method (maximum between-class variance) The threshold T; (D3) will Values ​​less than T are set to zero to obtain a one-dimensional high-resolution range-velocity vector. ; (D4) Transform the one-dimensional high-resolution range-velocity vector Converting it into a range-velocity two-dimensional matrix yields the target range and velocity information after sidelobe suppression.

2. A correlation-based frequency-agile radar sidelobe suppression system, characterized in that, Based on the frequency agile radar sidelobe suppression method described in claim 1, correlation-based frequency agile radar sidelobe suppression is achieved.

3. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, it implements correlation-based sidelobe suppression of frequency-agile radar based on the frequency-agile radar sidelobe suppression method of claim 1.

4. A computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, correlation-based sidelobe suppression of frequency-agile radar is achieved based on the frequency-agile radar sidelobe suppression method of claim 1.

Images