Radar broadband coherent signal DOA estimation method based on mirror symmetry

Abstract

The invention discloses a radar broadband coherent signal DOA (direction of arrival) estimation method based on mirror symmetry, which belongs to the field of array signal processing, and comprises the following steps: processing and sampling echo signals of each channel of a uniform linear array to obtain echo intermediate frequency signal data, and performing digital down-conversion processing to obtain echo signal baseband data; decomposing the data into data of a plurality of frequency points in a frequency domain, and calculating a data covariance matrix of each frequency point; performing decorrelation processing on the data covariance matrix of each frequency point, and performing characteristic decomposition to obtain a signal subspace and a noise subspace; performing spectrum peak search by using a mirror symmetry method to obtain a spatial spectrum on each frequency point, and fusing the spatial spectrums on the frequency points to obtain a final power spectrum; and carrying out spectrum peak search by using a mirror symmetry method to obtain a final DOA estimation result. According to the invention, the DOA angle of the broadband coherent signal can be accurately estimated, and the mirror symmetry method reduces the calculation amount of spectrum peak search in the algorithm.

Description

Technical Field

[0001] The present invention belongs to the technical field of radar signal processing, and relates to a DOA estimation technology for radar broadband coherent signals, and in particular to a DOA estimation method for radar broadband coherent signals based on mirror symmetry. Technical Background

[0002] Wideband signals carry rich frequency and time domain information, providing more target characteristics. Their wide frequency bands offer excellent anti-interference capabilities and exceptionally high resolution. Furthermore, wideband signals can meet diverse needs and possess considerable flexibility, leading to their widespread application in various fields. For wideband radar signals, the wide spectrum distribution causes the array flow matrix to vary, making traditional narrowband DOA techniques inapplicable. Furthermore, the increasingly complex spatial environment creates a large number of targets to be detected, while array element resources are limited. Therefore, strategies and methods that can detect more targets within limited array element resources are needed. Furthermore, wideband radar DOA estimation techniques generally suffer from high computational complexity, poor real-time performance, and poor robustness, making them of limited engineering value. Therefore, new methods are needed that can achieve high-precision, real-time, and rapid wideband DOA estimation that can process coherent signals. Summary of the Invention

[0003] The purpose of the present invention is to overcome the shortcomings of the prior art and propose a radar broadband coherent signal DOA estimation method based on mirror symmetry.

[0004] In order to achieve the above-mentioned purpose, the present invention adopts the following technical methods to achieve it.

[0005] The radar broadband coherent signal DOA estimation method based on mirror symmetry includes the following steps:

[0006] Step 1: Process and sample the echo signals of each channel of the uniform linear array, measure the echo signal bandwidth and perform polyphase extraction filtering to obtain the echo signal baseband data;

[0007] Step 2: Decompose the data into several frequency points in the frequency domain and calculate the data covariance matrix of each frequency point;

[0008] Step 3: Decorrelate the data covariance matrix of each frequency point and perform eigendecomposition. Use the broadband coherent signal source number estimation method based on broadband decomposition denoising and reconstruction to estimate the number of signal sources and obtain the signal subspace and noise subspace.

[0009] Step 4: Use the mirror symmetry method to search for spectrum peaks to obtain the spatial spectrum at each frequency point, and then fuse the spatial spectrum at each frequency point to obtain the final power spectrum;

[0010] Step 5: Use the mirror symmetry method to search for spectrum peaks and obtain the final DOA estimation result.

[0011] Compared with the prior art, the present invention has the following beneficial effects:

[0012] (1) The radar broadband coherent signal DOA estimation method based on mirror symmetry of the present invention has higher estimation stability and better estimation performance in a low signal-to-noise ratio environment when estimating the number of signal sources compared with the conventional signal source number estimation method.

[0013] (2) The radar broadband coherent signal DOA estimation method based on mirror symmetry of the present invention utilizes a broadband decorrelation method and can accurately estimate the target direction in a broadband coherent signal source environment.

[0014] (3) The radar broadband coherent signal DOA estimation method based on mirror symmetry of the present invention utilizes the mirror symmetry method to reduce the computational complexity of spectrum peak search, greatly saving computational time and resources, and having certain real-time performance.

[0015] (4) The radar broadband coherent signal DOA estimation method based on mirror symmetry of the present invention further improves the insufficient estimation accuracy of the mirror symmetry method in multi-target scenarios, thereby improving the estimation accuracy and estimation stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a flow chart of the steps of this method;

[0017] Figure 2 is the likelihood function value of the broadband coherent signal source number estimation method based on broadband decomposition, denoising and reconstruction under different criteria;

[0018] Figure 3 This is a performance comparison chart of broadband coherent signal source number estimation methods based on broadband decomposition, denoising and reconstruction under different criteria;

[0019] Figure 4 It is the spatial spectrum of the single-target mirror-symmetric ISSM-MUSIC algorithm;

[0020] Figure 5 This is a comparison chart of the estimated success probability of the single-target mirror-symmetric ISSM-MUSIC algorithm;

[0021] Figure 6 This is a comparison chart of the estimation errors of the single-target mirror-symmetric ISSM-MUSIC algorithm;

[0022] Figure 7 It is the spatial spectrum of the multi-objective mirror-symmetric ISSM-MUSIC algorithm;

[0023] Figure 8It is the spectrum of the secondary peak search space of the multi-target mirror-symmetric ISSM-MUSIC algorithm in the interval [27.8, 31.8];

[0024] Figure 9 It is the spectrum of the secondary peak search space of the multi-target mirror-symmetric ISSM-MUSIC algorithm in the interval [-31.8, -27.8];

[0025] Figure 10 This is a comparison chart of the estimated success probabilities of the multi-target mirror-symmetric ISSM-MUSIC algorithm;

[0026] Figure 11 This is a comparison chart of the estimation errors of the single-target mirror-symmetric ISSM-MUSIC algorithm. DETAILED DESCRIPTION

[0027] The present invention is further illustrated below with reference to the accompanying drawings and specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention. After reading the present invention, modifications of various equivalent forms of the present invention made by those skilled in the art all fall within the scope defined by the claims attached to this application.

[0028] like Figure 1 As shown, the present invention provides a radar broadband coherent signal DOA estimation method based on mirror symmetry, comprising the following steps:

[0029] Step 1: Process and sample the echo signals of each channel of the uniform linear array, measure the echo signal bandwidth and perform polyphase extraction filtering to obtain the echo signal baseband data;

[0030] The specific steps are as follows:

[0031] Step 1.1: Design a broadband coherent signal receiving module, including a receiving array antenna, a limiter module, a radio frequency bandpass filter (BPF0), a high-frequency low-noise amplifier (LNA), a mixer, an intermediate frequency bandpass filter (BPF1), an analog-to-digital converter (ADC), and a signal algorithm processing device (FPGA).

[0032] Step 1.2: According to the bandpass sampling theorem, the intermediate frequency sampling frequency is 2B, and the sampled data is sent to the FPGA for digital down-conversion processing to obtain the baseband data of the echo signal.

[0033] Step 2: Decompose the data into several frequency points in the frequency domain and calculate the data covariance matrix of each frequency point. The specific steps are as follows:

[0034] Step 2.1: The baseband echo signal data is Divide it into K segments in chronological order, namely Then, a J-point FFT transform is performed on the K sub-segment received signal data to divide it into J groups of mutually unrelated frequency domain sub-narrowbands, as shown below: Each row in the above matrix corresponds to each frequency point f j , j=1, 2, …, J K frequency point data, each column is obtained by performing J-point FFT transformation on K sub-segments in time.

[0035] Step 2.2: For each frequency point f j For example, the data covariance matrix on frequency is:

[0036] Step 3: Decorrelate the data covariance matrix of each frequency point and perform eigendecomposition. Use the broadband coherent signal source number estimation method based on broadband decomposition denoising and reconstruction to estimate the number of signal sources and obtain the signal subspace and noise subspace.

[0037] The specific steps are as follows:

[0038] Step 3.2: Perform eigendecomposition on the data covariance matrix to obtain M eigenvalues, and sort the eigenvalues ​​from large to small to obtain λ 1j ,λ 2j , ..., λ Nj , ..., λ Mj Among them, N eigenvalues ​​are eigenvalues ​​of the signal space, and the remaining MN eigenvalues ​​are eigenvalues ​​of the noise subspace, where M is the number of arrays and N is the number of sources.

[0039] Step 3.3: Estimate the number of broadband coherent sources and reconstruct the covariance matrix as follows: R′(f j )=R(f j )-λ Mj I Next, the reconstructed covariance matrix R′(f j ) to perform eigendecomposition, and then average the eigenvalues ​​at each frequency point of the corresponding signal: Obviously, for the covariance matrix R′(f j ) does not change the order of the average values. When the signals are independent, the number of broadband sources can be accurately estimated by using the AIC criterion or the MDL criterion on the eigenvalues ​​after denoising, reconstruction and averaging. The expression is as follows:

[0040] Step 3.4: Determine the signal subspace U based on the estimated number of sources S (f j ) and noise subspace U N (f j ).

[0041] Step 4: Use the mirror symmetry method to search for spectrum peaks to obtain the spatial spectrum at each frequency point, and then fuse the spatial spectrum at each frequency point to obtain the final power spectrum;

[0042] The specific steps are as follows:

[0043] Step 4.1: Use the mirror symmetry method to construct the spectrum peak search function at each frequency point as follows: in In the above formula, d is the array element spacing, f j is the center frequency of each frequency point, and c is the electromagnetic wave velocity. The angle search range is 0 to 90 degrees, and only half of the angle range needs to be searched.

[0044] Step 4.2: Fuse the spatial spectrum of each frequency point as follows Finally, the total spatial spectrum is obtained.

[0045] Step 5: Use the mirror symmetry method to search for spectrum peaks and obtain the final DOA estimation result.

[0046] The specific steps are as follows:

[0047] Step 5.1: Peak search: Search for an angle within the range of 0 to 90 degrees.

[0048] Step 5.2: Determine whether there is a mirror-symmetric signal source based on the following criteria: If the left side of the equation is much greater than 1, the angle is determined to be If the left side of the equation is close to 0, the angle is determined to be If the left side of the equation is close to 1, it is determined that there is a mirror-symmetric signal source, and proceed to step 3;

[0049] Step 5.3: Perform secondary peak search in the interval ρ and the mirror interval ρ′. The peak search formula is the basic MUSIC algorithm form, where Mirror range where θ mirIt is an angle value greater than 0, which is determined by the spatial spectrum resolution and the search step size, and the final angle is obtained.

[0050] In order to verify the practical effect of the method provided by the present invention, the following simulation experiments and analyses were performed based on the above scheme:

[0051] 1. Simulation conditions: The experimental simulation uses a uniform linear array with M = 8 array elements and the element spacing is set to the half wavelength corresponding to the highest frequency point, i.e. d = 0.5c / f j , where c is the speed of electromagnetic wave propagation; the incident signal source is a broadband linear frequency modulation signal and is coherent, the number of signal sources N = 2, the bandwidth B = 400MHz, the time width T = 10μs, and the initial phase The sampling frequency is 800 MHz, and the incident angles are -29.7° and 30° respectively; the number of time domain sampling snapshots is L=2048, and the number of frequency points is J=128.

[0052] 2. Simulation content and result analysis:

[0053] Simulation 1: The broadband coherent signal source number estimation method based on broadband decomposition denoising and reconstruction is used to estimate the number of signal sources. Under a signal-to-noise ratio of 10dB, the results of the broadband coherent signal source number estimation method based on the AIC and MDL principles are as follows: Figure 2 When the likelihood function takes its minimum value, the corresponding horizontal axis is the estimated number of sources. It can be seen that the broadband coherent source number estimation method based on the AIC and MDL principles estimates the number of sources to be 2, which meets the actual target set in the simulation and proves the effectiveness of this method.

[0054] Simulation 2: The number of sources is estimated using the broadband coherent signal source estimation method based on broadband decomposition denoising and reconstruction. The number of Monte Carlo simulations is 200. The relationship between the estimation performance of the broadband coherent signal source estimation method based on the AIC and MDL principles and the signal-to-noise ratio is shown in the figure below. Figure 3 As shown, it can be seen that the use of denoising and reconstruction methods can improve the stability of source estimation under low signal-to-noise ratio conditions.

[0055] Simulation 3: In the case of a single target, there is only one signal source, the angle is -30°, the signal-to-noise ratio is 10dB, and the spatial spectrum is obtained using the mirror symmetric spectrum peak search method as shown in the figure below: Figure 4 As shown. It can be seen that there is a spectrum peak at 30°. According to the judgment criteria, the calculated value is 2.114*10 -4 , which is close to 0, so the target angle is judged to be -30°, which is consistent with the real target set in the simulation, proving the effectiveness of this method.

[0056] Simulation 4: In the case of a single target, there is only one signal source, the angle is -30°, and the number of Monte Carlo simulations is 200. The performance comparison between the mirror symmetry method and the original method is as follows: Figure 5 and Figure 6 As shown in the figure, it can be seen that the estimation success probability of the mirror-symmetric MUSIC algorithm is slightly lower than that of the original MUSIC algorithm under extremely low signal-to-noise ratio. Apart from this, the estimation success probability and estimation error are roughly the same as those of the original MUSIC algorithm. However, the mirror-symmetric MUSIC algorithm only needs to search half the spectrum, which saves about 50% of the spectrum peak search computing resources compared to the MUSIC algorithm.

[0057] Simulation 5: In the case of multiple targets, as shown in the simulation conditions, the signal-to-noise ratio is 10dB, and the spatial spectrum is obtained using the mirror symmetric spectrum peak search method as shown in Figure 7 As shown in the figure, it can be seen that there is a spectrum peak at 29.8°. The value calculated according to the judgment criteria is 0.2806, which does not meet the conditions of being much greater than 1 and close to 0. It is judged that there is a mirror image source, and a secondary spectrum peak search is performed as shown in the figure. Figure 8 and Figure 9 As shown in the figure, the two target angles are finally obtained as -29.7° and 30°, which are consistent with the real targets set in the simulation, proving the effectiveness of this method.

[0058] Simulation 6: In the case of multiple targets, as shown in the simulation conditions, the number of Monte Carlo simulations is 200, and the performance of the multi-target mirror symmetry method is as follows: Figure 10 and Figure 11 As shown in the figure, the estimated success probabilities of the multi-target mirror-symmetric MUSIC algorithm and the MUSIC algorithm are roughly the same. Furthermore, it is clear that the single-target mirror-symmetric MUSIC algorithm has the largest estimation error in the multi-target scenario, while the remaining two algorithms have relatively similar estimation errors. Furthermore, due to the secondary peak search, the multi-target mirror-symmetric MUSIC algorithm requires more computational effort than the single-target mirror-symmetric MUSIC algorithm, but less computational effort than the traditional MUSIC algorithm.

Claims

1. A radar broadband coherent signal DOA estimation method based on mirror symmetry, characterized in that: The structure includes the following steps: Step 1: Process and sample the echo signals of each channel of the uniform linear array, measure the echo signal bandwidth and perform polyphase extraction filtering to obtain the echo signal baseband data; Step 2: Decompose the data into several frequency points in the frequency domain and calculate the data covariance matrix of each frequency point; Step 3: Decorrelate the data covariance matrix of each frequency point and perform eigendecomposition. Use the broadband coherent signal source number estimation method based on broadband decomposition denoising and reconstruction to estimate the number of signal sources and obtain the signal subspace and noise subspace. Step 4: Use the mirror symmetry method to search for spectrum peaks to obtain the spatial spectrum at each frequency point, and then fuse the spatial spectrum at each frequency point to obtain the final power spectrum; Step 5: Use the mirror symmetry method to search for spectrum peaks and obtain the final DOA estimation result.

2. The radar wideband coherent signal DOA estimation method based on mirror symmetry according to claim 1, characterized in that: The baseband signal acquisition method in step 1 is as follows: designing a broadband coherent signal receiving module, performing intermediate frequency sampling on the echo signal, and sending the sampled data to the FPGA. Furthermore, the step 1 specifically includes the following steps: Step 1.1: Design a broadband coherent signal receiving module, including a receiving array antenna, a limiter module, a radio frequency bandpass filter (BPF0), a high-frequency low-noise amplifier (LNA), a mixer, an intermediate frequency bandpass filter (BPF1), an analog-to-digital converter (ADC), and a signal algorithm processing device (FPGA). Step 1.2: According to the bandpass sampling theorem, the intermediate frequency sampling frequency is 2B, and the sampled data is sent to the FPGA for digital down-conversion processing to obtain the baseband data of the echo signal.

3. The radar wideband coherent signal DOA estimation method based on mirror symmetry according to claim 2, characterized in that: The step 2 specifically includes the following: Step 2.1: The baseband echo signal data is Divide it into K segments in chronological order, namely Then, a J-point FFT transform is performed on the K sub-segment received signal data to divide it into J groups of mutually unrelated frequency domain sub-narrowbands, as shown below: Each row in the above matrix corresponds to each frequency point f J , j=1, 2, …, J K frequency point data, each column is obtained by performing J-point FFT transformation on K sub-segments in time. Step 2.2: For each frequency point f J For example, the data covariance matrix on frequency is:

4. The radar wideband coherent signal DOA estimation method based on mirror symmetry according to claim 3, characterized in that: The step 3 specifically includes the following: Step 3.1: Decorrelate the data covariance matrix of each frequency point by using methods such as spatial smoothing or Toplize reconstruction; Step 3.2: Perform eigendecomposition on the data covariance matrix to obtain M eigenvalues, and sort the eigenvalues ​​from large to small to obtain λ 1j ,λ 2j ,…,λ Nj ,…,λ Mj Among them, N eigenvalues ​​are eigenvalues ​​of the signal space, and the remaining MN eigenvalues ​​are eigenvalues ​​of the noise subspace, where M is the number of arrays and N is the number of sources. Step 3.3: Estimate the number of broadband coherent sources and reconstruct the covariance matrix as follows: R′(f J )=R(f J )-λ Mj I Next, the reconstructed covariance matrix R′(f J ) to perform eigendecomposition, and then average the eigenvalues ​​at each frequency point of the corresponding signal: Obviously, for the covariance matrix R′(f j ) does not change the order of the average values. When the signals are independent, the number of broadband sources can be accurately estimated by using the AIC criterion or the MDL criterion on the eigenvalues ​​after denoising, reconstruction and averaging. The expression is as follows: Step 3.4: Determine the signal subspace U based on the estimated number of sources S (f J ) and the noise subspace U N (f J ).

5. The radar wideband coherent signal DOA estimation method based on mirror symmetry according to claim 4, characterized in that: The step 4 specifically includes the following: Step 4.1: Use the mirror symmetry method to construct the spectrum peak search function at each frequency point as follows: in In the above formula, d is the array element spacing, f j is the center frequency of each frequency point, and c is the electromagnetic wave velocity. The angle search range is 0 to 90 degrees, and only half of the angle range needs to be searched. Step 4.2: Fuse the spatial spectrum of each frequency point as follows Finally, the total spatial spectrum is obtained.

6. The radar wideband coherent signal DOA estimation method based on mirror symmetry according to claim 5, characterized in that: The step 5 specifically includes the following: Step 5.1: Peak search: Search for an angle within the range of 0 to 90 degrees. Step 5.2: Determine whether there is a mirror-symmetric signal source based on the following criteria: If the left side of the equation is much greater than 1, the angle is determined to be If the left side of the equation is close to 0, the angle is determined to be If the left side of the equation is close to 1, it is determined that there is a mirror-symmetric signal source, and proceed to step 3; Step 5.3: Perform secondary peak search in the interval ρ and the mirror interval ρ′. The peak search formula is the basic MUSIC algorithm form, where , mirror interval where θ mir It is an angle value greater than 0, which is determined by the spatial spectrum resolution and the search step size, and the final angle is obtained.

Images