Robust adaptive detection method for target in interference background

Abstract

The present application belongs to the technical field of wideband radar signal processing, and particularly relates to a target robust adaptive detection method under interference background. In view of the problem that the existing wideband radar range expansion target adaptive detector is difficult to consider the algorithm calculation complexity, CFAR characteristic and anti-interference performance, a target robust adaptive detection method with closed form under interference background is constructed based on the Gradient test criterion of two-step detector design procedure, while ensuring the CFAR characteristic, the calculation complexity, intelligent anti-interference and detection performance of the range expansion target adaptive detection algorithm and other aspects are considered, and the adaptive detection performance of the multi-channel wideband radar on the weak and small target under the complex interference environment is improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of wideband radar signal processing, and particularly relates to a target robust adaptive detection method under interference background. BACKGROUND

[0002] With the increase of radar bandwidth, the range resolution is further improved, and wideband radar is widely used in modern military and civilian fields such as anti-interference, counter-reconnaissance, accurate detection and imaging, high-precision tracking, target identification, etc. The adaptive detection of range expansion targets around wideband radar has become one of the hot issues in the radar field. Unlike the narrowband radar target echo signal which usually occupies only one range resolution unit, the energy of the wideband radar target scattering point may spread to the adjacent distance unit, showing as a "one-dimensional range image", forming a range expansion target. If the point target detection method is still used to detect the echo signal in a single distance unit, and the adjacent distance unit sampling is used to estimate the background clutter statistical characteristics, on the one hand, the strong scattering point energy of the range expansion target is easy to leak to the adjacent distance unit, resulting in signal pollution phenomenon, and further forming a shielding effect on the target signal of the single distance unit to be detected, so that the point target detection method is not good; on the other hand, in actual application, the radar detection faces complex electromagnetic environment, and there may be electronic countermeasure signals or various natural or man-made interference sources such as civil electromagnetic signals, in addition, the environment of the target is complex and changeable, so that the non-uniformity of the background clutter is enhanced, the number of pure clutter auxiliary data satisfying the independent and identically distributed is limited, and compared with the narrowband radar, this problem is particularly prominent in the wideband radar target detection scene, so that the existing range expansion target detection method is difficult to achieve ideal detection effect.

[0003] In addition, the multi-channel adaptive target detection under Gaussian clutter with unknown covariance matrix has always been a hot research topic. Generally, it is assumed that the clutter components in the observation data (also referred to as main data) from multiple distance units to be detected have the same clutter covariance matrix as the reference distance unit data (also referred to as auxiliary data) containing only pure clutter, and it is assumed that a set of auxiliary data without target signals is available to estimate the unknown clutter covariance matrix. In actual application scenarios, due to wavefront distortion, array calibration error and other reasons, the target signal steering vector may be mismatched. For radar search mode and other application scenarios, the detector needs to have strong robustness to mismatched signals. For the commonly used rank 1 signal model, the target steering vector is fixed and completely known, and it is difficult to cope with the aforementioned mismatch problem. If the overall data set composed of the main data of multiple distance units to be detected and the auxiliary data is used to model the target and interference signals in the subspace, and the GLRT criterion is used to construct the detection statistic, a subspace GLRT detector (abbreviated as S-GLRT-HE) for range expansion targets under uniform clutter plus structured interference can be obtained. The detector can obtain good detection performance, but the calculation process is relatively complex and is not convenient for solving. If the Rao test criterion is used, a subspace Rao detector (S-Rao-HE) for range expansion targets under uniform clutter plus structured interference can be obtained. Compared with the GLRT detector, the detection performance of the detector is improved in some set environments, but the robustness to mismatched signals is poor, and the calculation complexity of the detection statistic is high, which is not convenient for engineering implementation.

[0004] In view of the complex detection environment composed of receiver internal noise and external structured interference faced by the multi-channel wideband radar range expansion target adaptive detection, how to fully utilize the received data information, reasonably design the form of the range expansion target adaptive detector, maintain the constant false alarm rate (CFAR) characteristics, and effectively balance the algorithm calculation complexity and the detection performance is the key to improve the wideband radar detection capability in complex environment, and is one of the difficult problems faced by the multi-channel wideband radar range expansion target adaptive detection. SUMMARY

[0005] In order to overcome the problems in the prior art, the present application provides a target robust adaptive detection method in an interference background.

[0006] The technical solution of the present application to solve the above technical problems is as follows:

[0007] A target robust adaptive detection method in an interference background, comprising the following steps:

[0008] Step 1. Obtain the main data Z from K to-be-detected range cells; when the clutter covariance matrix M, the target coordinate matrix P, and the interference coordinate matrix Q are all unknown, take the partial derivative of the target parameter vector using the complex Gaussian probability density function of the main data Z under the target hypothesis, and combine the maximum likelihood estimation of the unknown target coordinate matrix under the target hypothesis and the maximum likelihood estimation of the unknown interference coordinate matrix under the non-target hypothesis to construct a two-step Gradient detection statistic for the range extended target under the condition of a known clutter covariance matrix;

[0009] Step 2. Obtain the auxiliary data Y from R reference range cells adjacent to the to-be-detected range cells, obtain the maximum likelihood estimation of the clutter covariance matrix based on the auxiliary data, substitute the maximum likelihood estimation of the clutter covariance matrix into the two-step Gradient detection statistic of the range extended target obtained in Step 1, replace the unknown clutter covariance matrix therein, and construct the detection statistic λ of the target robust adaptive detection method under the interference background;

[0010] Step 3. Set the detection threshold T according to the preset false alarm probability; compare the detection statistic λ with the detection threshold T. If λ ≥ T, it is determined that there is a range extended target in the current to-be-detected range cell, and the main data is not used as the auxiliary data for other subsequent to-be-detected range cells; otherwise, if λ < T, it is determined that there is no range extended target in the current to-be-detected range cell, and the main data is used as the auxiliary data for other subsequent to-be-detected range cells.

[0011] Furthermore, in Step 1, when the clutter covariance matrix M is known, the two-step Gradient detection statistic of the target robust adaptive detector under the interference background is:

[0012]

[0013] Where,

[0014]

[0015]

[0016]

[0017]

[0018] In the formula, the main data is represented as an N×K-dimensional complex matrix Z = [z1, z2,..., z K , and the N×1-dimensional received complex signal in the t-th to-be-detected range cell is represented as z t = s t + j t + c t (t = 1, 2,..., K), where the N×1-dimensional target complex signal vector st and N x 1 dimensional interference complex steering vector j t are all deterministic and can be expressed as s t = H p t and j t = J q t , respectively, where H and J are known N x p dimensional column full rank target signal subspace complex matrix and N x q dimensional interference signal subspace complex matrix, respectively, p x 1 dimensional complex vector p t and q x 1 dimensional complex vector q t represent unknown complex coordinate vectors of target signal and interference signal, respectively; spaces H and J are linearly independent, and an N x (p + q) dimensional column full rank augmented matrix B = [H J] is constructed, and p + q ≤ N, I m represents m x m dimensional identity matrix, and tr function represents trace of square matrix, superscripts (·) T and (·) H represent transpose and conjugate transpose, respectively, and · represents determinant of square matrix.

[0019] Further, in step 2, the clutter covariance matrix is derived using the complex Gaussian probability density function of the auxiliary data Y and is set to zero, and a clutter covariance matrix maximum likelihood estimate based on auxiliary data is obtained as sample covariance matrix S, that is,

[0020]

[0021] where the auxiliary data is represented as N x R dimensional complex matrix Y = [y1, y2,..., y R ], and R observation data y t (t = 1, 2,..., R) are obtained from R reference distance units adjacent to the distance unit to be detected, and it is assumed that y t (t = 1, 2,..., R) only contains pure clutter components, where the N x 1 dimensional complex vector y t (t = 1, 2,..., R) of the tth reference distance unit satisfies which are also independent and identically distributed among different distance units.

[0022] Further, in step 2, the clutter covariance matrix maximum likelihood estimate is brought into the distance spread target two-step Gradient detection statistic obtained in step 1, and the unknown clutter covariance matrix therein is replaced, and a detection statistic λ of a target robust adaptive detection method in an interference background is constructed:

[0023]

[0024] where,

[0025]

[0026] Compared with the prior art, the present application has the following technical effects:

[0027] 1) The target robust adaptive detection method under interference background is constructed, the detector has a closed form expression, and has a low calculation complexity, which is convenient for engineering implementation;

[0028] 2) For the interference environment with subspace structure, the target robust adaptive detection method under interference background of the present application can effectively suppress different intensity interference signals, and has good intelligent anti-interference performance;

[0029] 3) For the mismatched signal of the target signal steering vector, the target robust adaptive detection method under interference background of the present application can effectively detect the mismatched signal, and has strong detection robustness for the mismatched signal;

[0030] 4) The detection method of the present application balances the performance of algorithm calculation complexity, detection performance and mismatch robustness while maintaining the CFAR characteristic, and improves the adaptive detection performance of multi-channel wideband radar for weak and small targets and mismatched targets in complex environment;

[0031] 5) The method of the present application is suitable for partial non-wideband radar detection situations, for example, using low / middle resolution radar to detect large targets or to detect spatially adjacent point target groups (such as ship formation, aircraft formation, vehicle formation, etc.) moving at the same speed, and has good application prospect. BRIEF DESCRIPTION OF DRAWINGS

[0032] Figure 1 is a functional module diagram of the target robust adaptive detection method under interference background of the present application;

[0033] Figure 2 is a detection performance comparison diagram of the present application method and the existing detection method for matched signals;

[0034] Figure 3 is a detection performance comparison diagram of the present application method and the existing detection method for mismatched signals;

[0035] Figure 2 In the figure, N=8, K=15, R=16, p=3, q=2, false alarm probability P fa =10 -3 , interference clutter power ratio ICR=15dB;

[0036] Figure 3 In the figure, N=8, K=15, R=16, p=3, q=2, P fa =10 -3 , ICR=15dB, mismatch angle square value cos 2 φ=0.5. DETAILED DESCRIPTION

[0037] The principles and characteristics of the present application are described below in conjunction with the accompanying drawings, and the examples are only used to explain the present application and are not used to limit the scope of the present application.

[0038] In view of the problem that the existing wideband radar range expansion target adaptive detector is difficult to balance the algorithm calculation complexity, CFAR characteristics and anti-interference performance, a target robust adaptive detection method in an interference background is constructed based on a Gradient test criterion of a two-step detector design procedure, while ensuring the CFAR characteristics, the calculation complexity of the range expansion target adaptive detection algorithm, the intelligent anti-interference and detection performance and other aspects are considered, and the adaptive detection performance of the multi-channel wideband radar on a weak target in a complex interference environment is improved.

[0039] The target robust adaptive detection method in an interference background disclosed by the present application comprises the following steps:

[0040] Step 1. Obtain main data Z from K to-be-detected distance units; in the case that the clutter covariance matrix M, the target coordinate matrix P and the interference coordinate matrix Q are unknown, the partial derivative of the target parameter vector is solved by using the complex Gaussian probability density function of Z under the target hypothesis, and the distance expansion target two-step Gradient detection statistic under the known clutter covariance matrix condition is constructed by combining the maximum likelihood estimation of the unknown target coordinate matrix under the target hypothesis and the maximum likelihood estimation of the unknown interference coordinate matrix under the no-target hypothesis.

[0041] The specific steps comprise:

[0042] For a coherent radar system with a space-time joint channel number N, the binary hypothesis test problem of H0 and H1 is considered, wherein H0 hypothesis is that the target does not exist, and only pure clutter exists; H1 hypothesis is that the target, clutter and interference all exist.

[0043] Suppose that the target may occupy K to-be-detected distance units, under the assumption H1, the N×1-dimensional received complex signal in the tth to-be-detected distance unit is represented as z t =s t +j t +c t (t=1,2,...,K), wherein the N×1-dimensional target complex signal vector s t and the N×1-dimensional interference complex vector j t are assumed to be deterministic, and can be represented as s t =Ηp t and j t =Jq t respectively, H and J are known column full-rank N×p-dimensional target signal subspace complex matrix and N×q-dimensional interference signal subspace complex matrix respectively, p t and q are p×1-dimensional complex vector and q×1-dimensional complex vectort The unknown complex coordinate vectors representing the target signal and the interference signal are respectively represented by the main data, which can be represented as an N×K dimensional complex matrix Z = [z1, z2, ..., z...]. K Note that subspaces H and J are linearly independent. Construct an N×(p+q) dimensional column full-rank augmented matrix B = [HJ], satisfying p+q≤N. The N×1 dimensional clutter vector c in the t-th range cell to be detected... t It is a zero-mean complex circular Gaussian vector, denoted as: t = 1, 2, ..., K, and the clutter vectors between different distance cells are independent and identically distributed, where the N×N dimensional clutter covariance matrix M is an unknown Hermitian positive definite complex matrix.

[0044] Under hypotheses H0 and H1, the complex Gaussian probability density function (PDF) of the principal data Z can be expressed as follows:

[0045] f0(Z)=(π NK |M| K ) -1 exp{-tr[M -1 (Z-JQ)(Z-JQ) H ]} (1)

[0046] f1(Z)=(π NK |M| K ) -1 exp{-tr[M -1 (Z-BD)(Z-BD) H ]} (2)

[0047] Among them, the unknown coordinate matrix of the target subspace Unknown coordinate matrix of the interference subspace Superscript (·) T and(·) H represents the transpose and conjugate transpose, respectively; |·| represents the determinant of the square matrix; and the tr function represents taking the trace of the square matrix.

[0048] Given the clutter covariance matrix M, a robust adaptive detector for targets under interference backgrounds is constructed based on the Gradient test criterion of a two-step detector design procedure. Its two-step Gradient detection statistic for range-extended targets can be expressed as:

[0049]

[0050] in, Target parameter vector It is unknown, the interference parameter vector. It is unknown; This represents the maximum likelihood estimate of the objective parameter vector Θ under the H0 assumption; Θr0 Θ = (P, Q)T r the value under the H0 hypothesis, Θ = (P, Q)T r the maximum likelihood estimate under the H1 hypothesis; the vec function realizes the vectorization of a matrix.

[0051] the target parameter vector Θ by the complex Gaussian probability density function of the main data Z under the target hypothesis H0 r derivation, i.e., the derivative of formula (2) with respect to the target parameter vector Θ r partial derivation, and the following can be obtained:

[0052]

[0053] partial derivation of (1) with respect to the interference coordinate matrix Q, and setting the derivation result to zero, the maximum likelihood estimate of the interference coordinate matrix Q under the H0 hypothesis can be obtained as:

[0054]

[0055] wherein, I m represents an m x m dimensional unit matrix.

[0056]

[0057] wherein, I m represents an m x m dimensional unit matrix.

[0058] partial derivation of (2) with respect to D, and setting the derivation result to zero, the maximum likelihood estimate of D under the H1 hypothesis can be obtained as:

[0059]

[0060] wherein,

[0061] It is noted that the maximum likelihood estimate of the target coordinate matrix P under the H1 hypothesis (denoted as ) is the first p columns of , and thus the maximum likelihood estimate of the target parameter vector under the H1 hypothesis is:

[0062]

[0063] In addition, it is noted that the target signal does not exist under the H0 hypothesis, and thus Θ r0 = 0.

[0064] Substitute formula (6) and formula (8) into formula (3), and through multiple algebraic operations, when the clutter covariance matrix M is known, the two-step Gradient detection statistics of the target robust adaptive detector in the interference background is

[0065]

[0066] Step 2. Obtain the auxiliary data Y from R reference distance units adjacent to the distance unit to be detected, derive the clutter covariance matrix using the complex Gaussian probability density function of the auxiliary data Y and set to zero, obtain the maximum likelihood estimation of the clutter covariance matrix based on the auxiliary data, and bring the maximum likelihood estimation of the clutter covariance matrix into the distance spread target two-step Gradient detection statistics obtained in step 1, replace the unknown clutter covariance matrix therein, and construct the detection statistics λ of the target robust adaptive detection method in the interference background; the specific steps include:

[0067] In order to estimate the clutter covariance matrix M, R observation data y t (t=1, 2,..., R) are obtained from R reference distance units adjacent to the distance unit to be detected, it is assumed that y t (t=1, 2,..., R) only contains pure clutter components, the auxiliary data can be represented as an N×R complex matrix Y=[y1, y2,..., y R ], wherein the N×1 complex vector y t (t=1, 2,..., R) of the tth reference distance unit satisfies which is also independent and identically distributed among different distance units.

[0068] The complex Gaussian probability density function PDF of the auxiliary data Y can be represented as

[0069]

[0070] Derive the clutter covariance matrix M using formula (10) and set to zero, that is, let the corresponding derivative be zero, and the maximum likelihood estimation of the clutter covariance matrix M based on the auxiliary data is the sample covariance matrix S, that is,

[0071]

[0072] Substitute the maximum likelihood estimation of formula (11) into formula (9) to replace the unknown clutter covariance matrix M in formula (9), and the detection statistics of the target robust adaptive detector in the interference background is obtained.

[0073]

[0074] wherein,

[0075]

[0076] The method of the application constructs a target robust adaptive detector in a jamming background. As can be seen from equation (12), the proposed target robust adaptive detection method in a jamming background has a closed-form expression of a detection statistic, without the need for iterative operation. In addition, it is worth noting that compared with the S-GLRT-HE detector of a range expansion target, the algorithmic computational complexity of the target robust adaptive detection method in a jamming background is lower, and it has stronger detection robustness to the steering vector mismatch signal. In summary, the target robust adaptive detection method in a jamming background of the application can effectively balance the reasonable balance among the algorithmic computational complexity, mismatch robustness and detection performance while maintaining the CFAR characteristic.

[0077] Step 3. To maintain the CFAR characteristic of the detection method, a detection threshold T is set according to a preset false alarm probability; the detection statistic λ is compared with the detection threshold T, if λ≥T, it is determined that the current to-be-detected range cell has a range expansion target, and the main data is not used as auxiliary data for subsequent other to-be-detected range cells; otherwise, if λ<T, it is determined that the current to-be-detected range cell does not have a range expansion target, and the main data is used as auxiliary data for subsequent other to-be-detected range cells.

[0078] To verify the effectiveness of the method of the application, two embodiments are given in the specific implementation, the first embodiment is for a sea detection environment, and the second embodiment is for a ground detection environment.

[0079] Embodiment 1:

[0080] Referring to the drawings accompanying the specification Figure 1 , the specific implementation of embodiment 1 is divided into the following steps:

[0081] Step A1 uses a sea detection radar to irradiate a to-be-detected sea area, and obtains main data Z of K to-be-detected range cells; the main data Z is sent to a maximum likelihood estimation solving module under H0 hypothesis, a derivative module of a probability density function under H1 hypothesis and a maximum likelihood estimation solving module under H1 hypothesis; in the maximum likelihood estimation solving module under H0 hypothesis, the maximum likelihood estimation of Q under H0 hypothesis is obtained according to equation (5); in the derivative module of the probability density function under H1 hypothesis, the derivative result of the complex Gaussian probability density function of the main data Z under H1 hypothesis with respect to the target parameter vector Θ is obtained according to equation (4); in the maximum likelihood estimation solving module under H1 hypothesis, the maximum likelihood estimation of Θ under H1 hypothesis is obtained according to equation (8) r r ​​The results obtained by the maximum likelihood estimation module under the H0 hypothesis, the derivative module of the probability density function under the H1 hypothesis and the maximum likelihood estimation module under the H1 hypothesis are sent to the two-step Gradient detection statistic construction module under the condition of a known covariance matrix, a distance extended target two-step Gradient detection statistic under the condition of a known clutter covariance matrix is constructed according to formula (9), and is sent to the target robust adaptive detector construction module in an interference background.

[0082] It is worth noting that in step A1, the sea clutter component is modeled by using a complex Gaussian distribution, and meanwhile, considering that external interference in the actual sea environment may have an adverse effect on the adaptive detection of the range extended target, the external interference is also considered in the detector design process, and the subspace signal is used to model the interference, so as to reduce the mismatch influence of the interference signal. For the interference environment with a subspace structure, the distance extended target two-step Gradient intelligent fusion detection method can effectively suppress interference signals of different intensities, and has good intelligent anti-interference performance. For the mismatched target signal steering vector, the distance extended target two-step Gradient intelligent fusion detection method can effectively detect the mismatched signal, and has strong detection robustness for the mismatched signal.

[0083] In step A2, the non-target range around the sea area to be detected is irradiated by a radar, and R reference distance units containing only pure sea clutter auxiliary data Y are obtained; the auxiliary data Y is sent to the clutter covariance matrix maximum likelihood estimation module, the derivative of the clutter covariance matrix is obtained by using the complex Gaussian probability density function of Y and is set to zero, and the clutter covariance matrix maximum likelihood estimation based on the auxiliary data is obtained according to formula (11) is sent to the target robust adaptive detector construction module in an interference background. is brought into the distance extended target two-step Gradient detection statistic obtained in step A1, the unknown clutter covariance matrix is replaced, the detection statistic λ of the target robust adaptive detection method in an interference background is constructed according to formula (15), and λ is sent to the detection decision module.

[0084] It is worth noting that in step A2, compared with the S-GLRT-HE detector of the distance extended target, the algorithm complexity of the method is lower, and the detection robustness for the mismatched signal is stronger. In addition, the target robust adaptive detection method in an interference background has a closed-form expression, and compared with the existing distance extended target adaptive detection method, the performance balance of the algorithm complexity, the detection performance and the mismatch robustness is considered while the CFAR characteristic is maintained, and the adaptive detection capability of the multi-channel wideband radar for the weak and small target and the mismatched target on the sea surface in a complex electromagnetic environment is improved. ​

[0085] Step A3 sets the detection threshold T according to the preset false alarm probability. Specifically, the false alarm probability is set as P fa According to the Monte Carlo method, the detection threshold T is calculated according to 100 / P fa measured sea clutter data accumulated in the early stage. Considering the difficulty in obtaining sea clutter, if the actual amount of pure sea clutter measured data R is less than 100 / P fa , the missing 100 / P fa R clutter data can be obtained by simulation using a sea clutter simulation model, wherein the model parameters are reasonably estimated and set according to the pure sea clutter measured data obtained. Further, the detection statistic λ is compared with the detection threshold T. If λ≥T, it is determined that the current K distance units to be detected exist a range expansion target, and the main data is not used as auxiliary data for subsequent other distance units to be detected. Otherwise, if λ

[0086] The performance comparison results of the detector in the target-oriented vector matching environment are shown in the accompanying Figure 2 The results show that, compared with the existing range expansion target S-GLRT-HE, S-Rao-HE and other detectors, the detector of the present method has better detection performance in the matching environment.

[0087] Embodiment 2:

[0088] Referring to the accompanying Figure 1 , the specific implementation of embodiment 2 is divided into the following steps:

[0089] Step B1 uses the ground detection radar to irradiate the region to be detected, and obtains the main data Z of K distance units to be detected. The main data Z is sent to the maximum likelihood estimation solving module under H0 hypothesis, the derivative module of the probability density function under H1 hypothesis, and the maximum likelihood estimation solving module under H1 hypothesis. In the maximum likelihood estimation solving module under H0 hypothesis, the maximum likelihood estimation of Q under H0 hypothesis is obtained according to formula (5) In the derivative module of the probability density function under H1 hypothesis, the derivative result of the complex Gaussian probability density function of the main data Z under H1 hypothesis with respect to the target parameter vector Θ r is obtained according to formula (4); in the maximum likelihood estimation solving module under H1 hypothesis, the maximum likelihood estimation of Θ r under H1 hypothesis is obtained according to formula (8) The results obtained by the maximum likelihood estimation module under the H0 hypothesis, the derivative module of the probability density function under the H1 hypothesis and the maximum likelihood estimation module under the H1 hypothesis are sent to the two-step Gradient detection statistic construction module under the condition of a known covariance matrix, a distance extended target two-step Gradient detection statistic under the condition of a known clutter covariance matrix is constructed according to formula (9), and is sent to the target robust adaptive detector construction module in an interference background.

[0090] It is worth noting that in step B1, the clutter component is modeled by using a complex Gaussian distribution, and meanwhile, considering that the actual ground environment may be affected by external interference, which may adversely affect the adaptive detection of the range extended target, the external interference is also considered in the detector design process, and the subspace signal is used to model the interference, so as to reduce the mismatch influence of the interference signal. For the interference environment with subspace structure, the distance extended target two-step Gradient intelligent fusion detection method can effectively suppress different intensity interference signals, and has good intelligent anti-interference performance. For the mismatched target signal steering vector, the distance extended target two-step Gradient intelligent fusion detection method can effectively detect the mismatched signal, and has strong detection robustness for the mismatched signal.

[0091] In step B2, the non-target range around the region to be detected is irradiated by a radar, and R reference distance units containing only pure clutter auxiliary data Y are obtained; the auxiliary data Y is sent to the clutter covariance matrix maximum likelihood estimation module, the derivative of the clutter covariance matrix is obtained by using the complex Gaussian probability density function of Y and is set to zero, and the clutter covariance matrix maximum likelihood estimation based on the auxiliary data is obtained according to formula (11) The is sent to the target robust adaptive detector construction module in an interference background, and the is brought into the distance extended target two-step Gradient detection statistic obtained in step B1, the unknown clutter covariance matrix is replaced, the detection statistic λ of the target robust adaptive detection method in an interference background is constructed according to formula (15), and λ is sent to the detection decision module.

[0092] It is worth noting that in step B2, compared with the S-GLRT-HE detector of the distance extended target, the algorithm complexity of the method is lower, and the detection robustness for the mismatched signal is stronger. In addition, the target robust adaptive detection method in an interference background has a closed-form expression, compared with the existing distance extended target adaptive detection method, the performance balance of the algorithm complexity, the detection performance and the mismatch robustness is considered while the CFAR characteristic is maintained, and the adaptive detection ability of the multi-channel wideband radar for the ground weak target and the mismatched target in a complex electromagnetic environment is improved.

[0093] Step B3 sets the detection threshold T according to the preset false alarm probability: specifically, the false alarm probability is set as P fa According to the Monte Carlo method, the detection threshold T is calculated according to 100 / P fa The detection statistics λ is compared with the detection threshold T, if λ≥T, it is determined that the current K distance units to be detected exist a range expansion target, and the main data is not used as auxiliary data for subsequent other distance units to be detected; otherwise, if λ

[0094] The detector performance comparison results in the target-oriented vector mismatch environment are shown in the following table: Figure 3 The results show that, compared with the existing range expansion target S-GLRT-HE, S-Rao-HE and other detectors, the detector of the present method has better detection robustness in the mismatch environment.

[0095] The above description is only the preferred embodiment of the present application, and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for robust adaptive detection of a target in the presence of interference, characterized in that, The method comprises the following steps: Step 1. Obtain main data Z from K distance units to be detected; in the case that the clutter covariance matrix M, the target coordinate matrix P and the interference coordinate matrix Q are unknown, partial derivatives of the target parameter vector are solved by using the complex Gaussian probability density function of the main data Z under the target hypothesis, and the maximum likelihood estimation of the unknown target coordinate matrix under the target hypothesis and the maximum likelihood estimation of the unknown interference coordinate matrix under the non-target hypothesis are combined to construct a distance extended target two-step Gradient detection statistic under the condition of the known clutter covariance matrix; Step 2. Obtain auxiliary data Y from R reference distance units adjacent to the distance unit to be detected, obtain the maximum likelihood estimation of the clutter covariance matrix based on the auxiliary data, replace the unknown clutter covariance matrix in the distance extended target two-step Gradient detection statistic obtained in step 1 by bringing the maximum likelihood estimation of the clutter covariance matrix, and construct a detection statistic λ of the target robust adaptive detection method in the interference background; Step 3. Set a detection threshold T according to a preset false alarm probability; compare the detection statistic λ with the detection threshold T, if λ >= T, it is determined that the distance extended target exists in the current distance unit to be detected, and the main data is not used as auxiliary data for other distance units to be detected; otherwise, if λ < T, it is determined that the distance extended target does not exist in the current distance unit to be detected, and the main data is used as auxiliary data for other distance units to be detected.

2. The method of claim 1, wherein, In step 1, when the clutter covariance matrix M is known, the two-step Gradient detection statistic of the target robust adaptive detector in the interference background is: Wherein, In the formula, the master data is represented as an N×K dimensional complex matrix Z = [z1, z2, ..., z K The N×1 dimensional received complex signal in the t-th distance cell to be detected is represented as z. t =s t +j t +c t (t=1,2,...,K), where the N×1 dimensional target complex signal vector s t and N×1 dimensional interference complex vector j t All are assumed to be deterministic and can be represented as s respectively. t =Hp t and j t =Jq t H and J are the known full-rank N×p-dimensional target signal subspace complex matrix and N×q-dimensional interference signal subspace complex matrix, respectively, and p×1-dimensional complex vector p. t and q×1 dimensional complex vector q t Let H and J represent the unknown complex coordinate vectors of the target signal and the interference signal, respectively; the spaces H and J are linearly independent. Construct an N×(p+q) dimensional column full-rank augmented matrix B = [HJ], satisfying p+q≤N, I m Represents an m×m dimensional identity matrix, and the tr function represents taking the trace of the square matrix, with the superscript (·). T and(·) H represents the transpose and conjugate transpose, respectively, and |·| represents the determinant of the square matrix.

3. The method of claim 2, wherein, In step 2, the maximum likelihood estimation of the clutter covariance matrix based on the auxiliary data is obtained by using the complex Gaussian probability density function of the auxiliary data Y to derive the clutter covariance matrix and set it to zero, that is, where the auxiliary data is represented as an N x R dimensional complex matrix Y = [y1, y2,..., yR], R observation data y R ] are obtained from R reference range cells adjacent to the range cell to be detected, it is assumed that y t (t = 1, 2,..., R) only contains pure clutter components, where the N x 1 dimensional complex vector y t (t = 1, 2,..., R) of the tth reference range cell satisfies t (t = 1, 2,..., R) are independent and identically distributed (i.i.d.). which are also i.i.d. among different range cells.

4. The method of claim 3, wherein, In step 2, the maximum likelihood estimation of the clutter covariance matrix is brought into the distance extended target two-step Gradient detection statistic obtained in step 1 to replace the unknown clutter covariance matrix therein, and the detection statistic λ of the target robust adaptive detection method in the interference background is constructed: Wherein,

Images