Target parameter estimation method for spatial stepped-frequency radar in pulse noise environment

Abstract

The invention discloses a method for estimating target parameters of a spatial stepped frequency radar in an impulse noise environment. The method includes establishing an impulse noise signal model in an impulse noise environment; preprocessing a received first signal based on the impulse noise signal model to obtain a second signal; generating a covariance matrix according to the second signal; adding a median differential correlation entropy as a weighting factor to the covariance matrix when generating the covariance matrix; solving the covariance matrix using a MUSIC solution method to obtain position parameters and angle parameters of a detected target; the invention adopts the median differential correlation entropy to suppress impulse noise, which can not only effectively solve the problem of unbounded output covariance matrix in an impulse noise environment, but also has a good inhibitory effect on abnormal mutation values ​​in a received signal in a strong impulse noise environment, and can ensure the estimation accuracy of target parameters in a subsequent impulse noise environment.

Description

Technical Field

[0001] The invention belongs to the technical field of ground penetrating radar target parameter estimation, and in particular relates to a spatial stepped frequency radar target parameter estimation method in an impulse noise environment. Background Art

[0002] Space stepped frequency radar is an ultra-wideband ground penetrating radar that combines a multi-transmitter and multi-receiver array system with a stepped frequency signal. Due to the presence of a large amount of non-Gaussian pulse noise in the actual environment, such as environmental noise, machine cycle noise, radar reflection clutter, explosion, discharge, etc., many abnormal pulse values ​​appear in the radar receiving signal. In addition, non-Gaussian pulse noise conforms to the α-stable distribution characteristics and does not have second-order and higher moment characteristics, which makes the resolution performance of the existing target parameter estimation method based on second-order and higher-order statistical characteristics drop sharply or even fail, posing a great threat to the reliability of the actual application of space stepped frequency radar.

[0003] In recent years, many researchers have conducted in-depth research and discussion on radar target parameter estimation under the background of α-stable distribution noise, mainly focusing on two types of methods: fractional lower-order statistics and related entropy function. In 2001, Mendel et al. proposed the fractional lower-order moment algorithm (FLOM) in "A subspace-based direction finding algorithm using fractional lower order statistics", which significantly improved the suppression ability of weak pulse noise, but the algorithm was too large in computation due to the influence of nonlinear operators; in 2010, Belkacemi et al. proposed the phase fractional lower-order moment (PFLOM) algorithm in "Robust subspace-based algorithms for joint angle / Dopplerestimation in non-Gaussian clutter", which has good suppression ability for strong pulse noise. However, all of the above methods require prior information of the pulse noise signal, and the suppression performance of α-stable distribution noise is seriously reduced when the signal-to-noise ratio is low and the characteristic index is small.

[0004] In 2014, Zhang et al. proposed a Correntropy Based Correlation (CRCO) algorithm in "A novel correntropy based DOA estimation algorithm in impulsive noise environments". The correlation entropy function is used as a weight coefficient to multiply the covariance matrix of the received signal, thereby ensuring the orthogonality of the signal subspace and the noise subspace, and achieving accurate estimation of coherent sources under impulse noise background. In 2020, Chen et al. proposed a Correntropy Based Covariance Matrix (CBCM) algorithm in "Correntropy-based DOA estimation algorithm under impulsive noise environments". The correlation entropy function is used to reconstruct the covariance matrix of the received signal and combined with the MUSIC algorithm to effectively improve the DOA estimation accuracy of coherent sources under impulse noise background.

[0005] However, both the fractional low-order statistics method and the correlation entropy function method have poor noise suppression performance under low signal-to-noise ratio conditions. Summary of the invention

[0006] The purpose of the present invention is to provide a method for estimating target parameters of a space stepped frequency radar in an impulse noise environment, so as to improve the suppression performance of the space stepped frequency radar on impulse noise in an impulse noise environment with a low signal-to-noise ratio, and realize accurate estimation of target distance and angle parameters.

[0007] The present invention adopts the following technical solution: a method for estimating target parameters of a spatial stepped frequency radar in an impulse noise environment, comprising the following steps:

[0008] Establish an impulse noise signal model in an impulse noise environment;

[0009] Preprocessing the received first signal based on the impulse noise signal model to obtain a second signal;

[0010] Generating a covariance matrix of the second signal according to the second signal; wherein, when generating the covariance matrix, the median difference correlation entropy is added to the covariance matrix as a weighting factor;

[0011] The MUSIC solution method is used to solve the covariance matrix to obtain the position parameters and angle parameters of the detected target.

[0012] Furthermore, the covariance matrix is ​​solved by using the MUSIC solution method, including:

[0013] Perform eigenvalue decomposition on the covariance matrix to obtain the signal subspace and noise subspace;

[0014] According to the orthogonal characteristics of the signal subspace and the noise subspace, the distance parameter and angle parameter of the pth detected target are calculated.

[0015] Furthermore, the (i,j)th element in the covariance matrix is:

[0016]

[0017] Among them, R MCBCM (i,j) is the (i,j)th element in the covariance matrix, L is the number of snapshots of the second signal, l is the sequence number of the snapshot, med(·) is the median operation, y i is the row vector of the i-th row in the second signal, for y i The conjugate of j is the column vector of the jth column in the second signal, for y j The conjugate of , μ1 is the suppression parameter, and δ is the kernel length of the kernel function.

[0018] Furthermore, the second signal is:

[0019]

[0020] Among them, y l is the second signal of the lth snapshot, is the total number of detected targets, α p,l is the complex amplitude of the echo signal of the pth target at the lth snapshot, a r (θ p ) is the receiving guidance vector corresponding to the pth target, θ p is the angle of the incoming wave of the pth target, is the Kronecker product, a t (r p ,θ p ) is the launch guidance vector corresponding to the pth target, r p is the distance from the pth target to the array antenna, is the impulse noise vector at time t, with a dimension of NM×L, where M is the number of transmitting array elements and N is the number of receiving array elements.

[0021] Furthermore, the impulse noise signal model is:

[0022]

[0023] in, is the impulse noise vector at time t, j is the imaginary sign, μ is the position parameter, t is the time, γ is the scale coefficient, α represents the characteristic exponent, β is the symmetry parameter, and sgn(t) is the sign function.

[0024] Furthermore, the eigenvalue decomposition of the covariance matrix is ​​specifically:

[0025]

[0026] Among them, R MCBCM is the covariance matrix, U S is the signal subspace, Σ S is the eigenvector corresponding to the signal subspace, U N is the noise subspace, Σ N is the eigenvector corresponding to the pseudo noise subspace.

[0027] Furthermore, the method for calculating the distance parameter and angle parameter of the pth detected target is:

[0028]

[0029] in, are the distance and angle parameters of the pth detected target,

[0030] Another technical solution of the present invention: a device for estimating target parameters of a space stepped frequency radar in an impulse noise environment, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the method for estimating target parameters of a space stepped frequency radar in an impulse noise environment is implemented.

[0031] The beneficial effects of the present invention are as follows: the present invention adopts the median differential correlation entropy to suppress impulse noise, which can not only effectively solve the problem of unbounded output covariance matrix in the impulse noise environment, but also has a good inhibitory effect on abnormal mutation values ​​in the received signal in the strong impulse noise environment, and can ensure the estimation accuracy of the target parameters in the subsequent impulse noise environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is a pulse noise simulation diagram established by selecting α stable distribution in an embodiment of the present invention;

[0033] Figure 2 is a preprocessing diagram of the mth array element of the spatial stepped frequency radar receiving signal in an embodiment of the present invention;

[0034] Figure 3 This is the result of simulating the distance and angle two-dimensional power spectrum estimation using different methods when the generalized signal-to-noise ratio is 15dB, the number of snapshots is 100, and the characteristic index α=1.2;

[0035] Figure 4 This is the result of simulating the distance and angle two-dimensional power spectrum estimation using different methods when the generalized signal-to-noise ratio is 5dB, the number of snapshots is 100, and the characteristic index α=1.2;

[0036] Figure 5 It is a curve of mean square error of distance using different methods versus generalized signal-to-noise ratio;

[0037] Figure 6 It is a curve diagram showing the correct probability of estimation with the generalized signal-to-noise ratio using different methods and angles;

[0038] Figure 7 It is a graph showing the mean square error of the distance using different methods versus the characteristic index;

[0039] Figure 8 It is a curve chart showing the change of correct probability with characteristic index using different methods and angles. DETAILED DESCRIPTION

[0040] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0041] The invention discloses a method for estimating target parameters of a space stepped frequency radar in an impulse noise environment, comprising the following steps: establishing an impulse noise signal model in the impulse noise environment; preprocessing a received first signal based on the impulse noise signal model to obtain a second signal; generating a covariance matrix according to the second signal; wherein, when generating the covariance matrix, adding the median difference correlation entropy as a weighting factor into the covariance matrix; and solving the covariance matrix using a MUSIC solution method to obtain position parameters and angle parameters of the detection target.

[0042] The present invention adopts median differential correlation entropy to suppress impulse noise, which can not only effectively solve the problem of unbounded output covariance matrix in impulse noise environment, but also has a good inhibitory effect on abnormal mutation values ​​in received signals in strong impulse noise environment, and can ensure the estimation accuracy of target parameters in subsequent impulse noise environment.

[0043] The invention relates to two-dimensional joint estimation of target distance and angle parameters of a space stepped frequency radar in an impulse noise environment, and specifically to a median difference correlation entropy method based on a covariance matrix, which can effectively suppress strong impulse noise signals and realize high-precision estimation of target distance and angle parameters in an impulse noise environment with a low signal-to-noise ratio and few snapshots. When the invention detects a target using a space stepped frequency radar based on a multiple-transmit and multiple-receive system, the transmitted stepped frequency signal increases the signal bandwidth and improves the distance resolution, and the receiving end adopts aperture synthesis to improve the angle resolution.

[0044] In one embodiment, solving the covariance matrix using the MUSIC solution method includes: performing eigenvalue decomposition on the covariance matrix to obtain a signal subspace and a noise subspace; and calculating the distance parameter and angle parameter of the pth detected target according to the orthogonal characteristics of the signal subspace and the noise subspace.

[0045] Specifically, the search space is divided into a two-dimensional grid with K angle search units and P distance search units. The covariance matrix based on the median difference correlation entropy is decomposed by eigenvalue, and the covariance matrix based on the median difference correlation entropy under L snapshots can be decomposed into:

[0046]

[0047] Among them, R MCBCM is the covariance matrix, U S is the signal subspace, Σ S is the eigenvector corresponding to the signal subspace, U N is the noise subspace, Σ N is the eigenvector corresponding to the pseudo noise subspace.

[0048] Next, according to the orthogonal characteristics of the signal subspace and the noise subspace, a new two-dimensional search guidance vector is constructed. By using the orthogonal characteristics to construct a two-dimensional search guidance vector, the spatial power spectrum can be obtained. That is, the method for calculating the distance parameter and angle parameter of the pth detection target is:

[0049]

[0050] in, are the distance and angle parameters of the pth detected target, is the guiding vector.

[0051] In the embodiment of the present invention, since the median has the ability to characterize the general level of a set of data and is not easily affected by the maximum or minimum values ​​in the received data, the median difference idea is used to construct a new correlation entropy operator, and the Gaussian kernel function is selected as the kernel function, and then the (i, j)th element in the covariance matrix is:

[0052]

[0053] Among them, R MCBCM (i,j) is the (i,j)th element in the covariance matrix, L is the number of snapshots of the second signal, l is the sequence number of the snapshot, med(·) is the median operation, y i is the row vector of the i-th row in the second signal, for y i The conjugate of j is the column vector of the jth column in the second signal, for yj The conjugate of , μ1 is the suppression parameter, and δ is the kernel length of the kernel function.

[0054] By measuring the deviation between the sampling value and the median to construct a correlation entropy function, and importing the median difference correlation entropy as a weighting factor into the covariance matrix of the received signal, the effect of enhancing the suppression of impulse noise can be achieved.

[0055] In the embodiment of the present invention, a spatial stepped frequency radar echo signal model is first established in an α-stable distribution noise environment, such as Figure 1 As shown in the figure, it is a simulation diagram of impulse noise when different α values ​​are taken. Since non-Gaussian impulse noise has a heavy tail statistical characteristic, it completely obeys the α stable distribution, that is, the impulse noise signal model is:

[0056]

[0057] in, is the impulse noise vector at time t, j is the imaginary number symbol,

[0058] α represents the characteristic index. When 1<α≤2, its value determines the impact strength and tail thickness of the α-stable distribution. As the α value decreases, the impact of the α-stable distribution becomes stronger and the corresponding tail becomes thicker. On the contrary, as the α value increases, the impact of the α-stable distribution becomes weaker and the corresponding tail becomes thinner. Especially when α is closer to 2, its distribution characteristics are closer to the traditional Gaussian distribution characteristics. When α=2, it is a Gaussian random distribution. Therefore, the α-stable distribution can accurately describe the non-Gaussian impulse noise signal.

[0059] μ is the location parameter, which is the center of the probability density function of the α-stable distribution. μ is the mean of the α-stable distribution. When 0<α≤1, μ is the median of the α-stable distribution. t is the time. γ is the scale coefficient, which determines the degree of dispersion of the stable distribution variable α from the termination or mean. Its characteristics are similar to the variance of the Gaussian distribution. β is the symmetry parameter, which mainly determines the degree of tilt and distortion of the α-stable distribution. When β=0, the probability density function of the α-stable distribution presents symmetric characteristics, that is, symmetric α-stable distribution. sgn(t) is the sign function. The value range of each parameter is 0<α≤2, -1<β≤1, γ>0, -∞<μ<∞.

[0060] like Figure 2 As shown, the received signal (i.e., the first signal) is preprocessed by mixing, sampling, matching separation, and low-pass filtering to obtain a second signal:

[0061]

[0062] Among them, y l is the second signal of the lth snapshot, is the total number of detected targets, αp,l is the complex amplitude of the echo signal of the pth target at the lth snapshot, a r (θ p ) is the receiving guidance vector corresponding to the pth target, θ p is the angle of the incoming wave of the pth target, is the Kronecker product, a t (r p ,θ p ) is the launch guidance vector corresponding to the pth target, r p is the distance from the pth target to the array antenna, is the impulse noise vector at time t, with a dimension of NM×L, where M is the number of transmitting array elements and N is the number of receiving array elements.

[0063] Among them, after preprocessing, the steering vector a t (r p ,θ p ) and a r (θ p ) is expressed as:

[0064]

[0065] Among them, λ=c / f0 is the wavelength of the reference array element, Δf is the step frequency increment, c is the speed of light, d is the array element spacing, λ represents the wavelength of the electromagnetic wave, and f0 is the initial wavelength.

[0066] In order to verify the accuracy of the method of the present invention, the following simulation verification was carried out.

[0067] 1. Simulation data:

[0068] A uniform array with co-location of transmitter and receiver is used. The number of transmitting array elements and receiving array elements are N = M = 9, and the array element spacing is d. t =d r =d=λ / 2, f0=10GHz, the step frequency increment Δf is 2MHz, the sampling frequency f s =1 / T e =18MHz, then the equivalent transmission signal bandwidth B w =NΔf=18MHz, the corresponding distance resolution is Δr=c / (2B w )=8.333m,speed of light c=3×10 8 m / s.

[0069] Assume that the specific positions of the targets are (r1, θ1) = (9m, 2°), (r2, θ2) = (20m, 5°), (r3, θ3) = (31m, 15°), and the corresponding powers are 10dB, 8dB, and 6dB, respectively. Set the distance search interval to [0:0.2:40]m, and the angle search interval to [0°:0.5°:20°]. The α-stable distribution is used to establish the impulse noise model. Since the α-stable distribution does not have a finite variance, the conventional signal-to-noise ratio (SNR) cannot be applied. Therefore, the present invention uses the generalized signal-to-noise ratio (GSNR) to measure the signal strength of the α-stable distribution noise. The specific expression of GSNR is:

[0070]

[0071] Among them, δ s is the average signal power, N a is the number of samples.

[0072] 2. Simulation content and results.

[0073] In simulation 1, assuming that the generalized signal-to-noise ratio is 5 dB and 15 dB, the number of snapshots is 300, and the characteristic index α is 1.2, the impulse noise suppression performance of the FLOM, PFLOM, CRCO, CBCM algorithms and the method of the present invention is compared, and then the 2D-MUSIC algorithm is used to obtain the two-dimensional power spectrum estimation results of the target distance and angle. Figure 3 and Figure 4 shown.

[0074] Figure 3 is a two-dimensional power spectrum estimation result diagram of each method and the method of the present invention when GSNR=5dB, Figure 3 It can be seen that under low signal-to-noise ratio, the resolution performance of each method is poor, while the method of the present invention can estimate the target distance and angle parameters and has a high performance in suppressing impulse noise.

[0075] Figure 4 When SNR=15dB, the two-dimensional power spectrum estimation results of each method and the method of the present invention are shown in the figure, where the black circle is the location of the real target. Figure 5 It can be seen that under high signal-to-noise ratio, compared with other methods, the method of the present invention can effectively suppress impulse noise and achieve high-precision estimation of target distance and angle parameters.

[0076] In simulation 2, different experimental conditions are set to evaluate the pulse noise suppression performance of each algorithm and the method of the present invention. 200 Monte Carlo experiments are performed under each condition. The mean square error and estimation accuracy are selected as evaluation criteria, and the root mean square error and estimation accuracy curves of each method under different experimental conditions are plotted.

[0077] Depend on Figure 5 and Figure 6 It can be seen that the distance and angle mean square errors of the method (MCBCM) of the present invention are lower than those of other methods under different generalized signal-to-noise ratios. When the GSNR is small, the distance and angle mean square errors of other algorithms increase significantly, and the estimation accuracy is low. However, the estimation accuracy of the method of the present invention is as high as more than 90%. It can be concluded that the method of the present invention has good suppression ability for strong impulse noise under low signal-to-noise ratio conditions.

[0078] Depend on Figure 7 and Figure 8 It can be seen that the mean square error of distance and angle of the method of the present invention is lower than that of other methods under different characteristic indexes. The increase of characteristic index value corresponds to the gradual change of α stable distribution noise from extremely strong impulse noise to Gaussian white noise, and the mean square error of each algorithm decreases accordingly, and the estimation accuracy increases accordingly. In the entire value range of characteristic index, the mean square error of distance and angle of the method of the present invention always remains the smallest, and the estimation accuracy is the highest, indicating that the ability of the method of the present invention to suppress impulse noise is significantly better than other algorithms, and it has good applicability to both weak impulse noise and strong impulse noise.

[0079] In summary, Figure 3-Figure 8 It is shown that under low signal-to-noise ratio conditions, the FLOM, PFLOM, CRCO, and CBCM methods have poor impulse noise suppression performance, resulting in a serious decrease in the accuracy of estimated target distance and angle. The method of the present invention can effectively suppress the influence of impulse noise on the received signal through median difference, and obtains good distance and angle estimation accuracy through the 2D-MUSIC algorithm, providing an effective impulse noise suppression method for the practical application of space stepped frequency radar.

[0080] The present invention also discloses a device for estimating target parameters of a space stepped frequency radar in an impulse noise environment, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the method for estimating target parameters of a space stepped frequency radar in an impulse noise environment is implemented.

[0081] The above-mentioned device can be a computing device such as a desktop computer, a notebook, a PDA, and a cloud server. The device may include but is not limited to a processor and a memory. Those skilled in the art will appreciate that the device may include more or fewer components, or a combination of certain components, or different components, for example, it may also include input and output devices, network access devices, etc.

[0082] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0083] In some embodiments, the memory may be an internal storage unit of the device, such as a hard disk or memory of the device. In other embodiments, the memory may also be an external storage device of the device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the device. Further, the memory may also include both an internal storage unit of the device and an external storage device. The memory is used to store an operating system, an application program, a boot loader (BootLoader), data, and other programs, such as the program code of the computer program, etc. The memory may also be used to temporarily store data that has been output or is to be output.

[0084] It should be noted that the specific content of the above-mentioned device is based on the same concept as the embodiment of the method of the present invention. Its specific functions and technical effects can be found in the method embodiment part, and will not be repeated here.

Claims

1. A method for estimating target parameters of a spatial stepped frequency radar in an impulse noise environment, characterized in that: The following steps are involved: Establish an impulse noise signal model in an impulse noise environment; Based on the impulse noise signal model, preprocessing the received first signal to obtain a second signal; Generating a covariance matrix of the second signal according to the second signal; wherein, when generating the covariance matrix, the median difference correlation entropy is added to the covariance matrix as a weighting factor; The covariance matrix is ​​solved by using the MUSIC solution method to obtain the position parameters and angle parameters of the detection target.

2. The method for estimating spatial stepped frequency radar target parameters in an impulse noise environment as claimed in claim 1, characterized in that: Solving the covariance matrix using the MUSIC solution method includes: Performing eigenvalue decomposition on the covariance matrix to obtain a signal subspace and a noise subspace; According to the orthogonal characteristics of the signal subspace and the noise subspace, the distance parameter and angle parameter of the pth detected target are calculated.

3. The method for estimating spatial stepped frequency radar target parameters in an impulse noise environment as claimed in claim 2, characterized in that: The (i, j)th element in the covariance matrix is: Among them, R MCBCM (i, j) is the (i, j)th element in the covariance matrix, L is the number of snapshots of the second signal, l is the sequence number of the snapshot, med(·) is the median operation, y i is the row vector of the i-th row in the second signal, for y i The conjugate of j is the column vector of the jth column in the second signal, for y j The conjugate of , μ1 is the suppression parameter, and δ is the kernel length of the kernel function.

4. The method for estimating spatial stepped frequency radar target parameters in an impulse noise environment as claimed in claim 3, characterized in that: The second signal is: Among them, y l is the second signal of the lth snapshot, is the total number of detected targets, α p,l is the complex amplitude of the echo signal of the pth target at the lth snapshot, a r (θ p ) is the receiving guidance vector corresponding to the pth target, θ p is the angle of the incoming wave of the pth target, is the Kronecker product, a t (r p ,θ p ) is the launch guidance vector corresponding to the pth target, r p is the distance from the pth target to the array antenna, is the impulse noise vector at time t, with a dimension of NM×L, where M is the number of transmitting array elements and N is the number of receiving array elements.

5. The method for estimating spatial stepped frequency radar target parameters in an impulse noise environment as claimed in claim 4, characterized in that: The impulse noise signal model is: in, is the impulse noise vector at time t, j is the imaginary sign, μ is the position parameter, t is the time, γ is the scale coefficient, α represents the characteristic exponent, β is the symmetry parameter, and sgn(t) is the sign function.

6. The method for estimating spatial stepped frequency radar target parameters in an impulse noise environment according to any one of claims 2 to 5, characterized in that: The specific steps of performing eigenvalue decomposition on the covariance matrix are as follows: Among them, R MCBCM is the covariance matrix, U S is the signal subspace, Σ S is the eigenvector corresponding to the signal subspace, U N is the noise subspace, Σ N is the eigenvector corresponding to the pseudo noise subspace.

7. The method for estimating target parameters of a spatial stepped frequency radar in an impulse noise environment as claimed in claim 6, characterized in that: The method for calculating the distance parameter and angle parameter of the pth detected target is: in, are the distance and angle parameters of the pth detected target, 8. A device for estimating target parameters of a space stepped frequency radar in an impulse noise environment, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method for estimating target parameters of a space stepped frequency radar in an impulse noise environment as described in any one of claims 1 to 7 when executing the computer program.

Images