Radar distributed target detection method and system during fixed interference steering vector mismatching

Abstract

The invention belongs to the field of radar target detection, and discloses a radar distributed target detection method and system during fixed interference steering vector mismatching. In order to solve the problem that in the prior art, when a fixed interference steering vector is mismatched, reliable detection performance cannot be guaranteed, the invention provides a radar distributed target detection method when the fixed interference steering vector is mismatched, and the radar distributed target detection method comprises the following steps: firstly, estimating a noise covariance matrix based on auxiliary data; then estimating an amplitude vector of fixed interference and a steering vector coordinate vector, a target coordinate vector and a noise power mismatching amount thereof based on to-be-detected data, further calculating a detection statistic based on a generalized likelihood ratio criterion, determining a detection threshold according to a false alarm probability, and finally comparing the detection statistic with the detection threshold to obtain a detection result. If the former is large, the target exists, and if the latter is large, the target does not exist. According to the method, the reliable detection performance can be ensured under the condition of mismatching of the fixed interference steering vector.

Description

Technical Field

[0001] The present invention belongs to but is not limited to the field of radar target detection technology, and in particular relates to a radar distributed target detection method and system when a fixed interference guidance vector is mismatched. Background Art

[0002] As one of the basic functions of radar, adaptive target detection continues to receive extensive attention and research. When detecting targets of interest, radars are also affected by factors such as noise, clutter and interference. Among them, the target of interest may occupy multiple range units under the observation of high-resolution radar, and its echo can be described by a distributed model. Noise and clutter can be uniformly modeled as zero-mean Gaussian colored noise, and the covariance matrix of the colored noise can be calculated by obtaining auxiliary data containing only independent and identically distributed noise from adjacent range resolution units. Interference may come from a jammer in a fixed direction. After entering the radar receiver, the fixed interference may disrupt the radar's correct detection of the target of interest. The location of the jammer can be known in advance through a series of interference identification techniques, so current studies all assume that the guidance information of the fixed interference is known a priori.

[0003] Aiming at the problem of radar distributed target detection with a priori known guidance information of fixed interference, Francesco Bandiera et al. designed a detector based on the generalized likelihood ratio test (GLRT) in 2007, and Liu Weijian et al. designed detectors based on the Rao criterion and the Wald criterion in 2015 and 2017, respectively. These detectors have different performance advantages and disadvantages in different scenarios: the Rao detector performs better when the sum of the signal subspace and interference subspace dimensions is close to the data model dimension; the Wald detector performs better when the data model dimension is large and the signal subspace dimension is larger than the interference subspace dimension; the GLRT detector performs better in other cases.

[0004] Existing detectors guarantee reliable detection performance when the fixed interference guidance information is accurately known. However, in increasingly complex real environments, due to factors such as array calibration errors and interference identification errors, mismatches in the guidance vectors of fixed interferences are common, that is, there is a deviation between the assumed guidance vector and the actual guidance vector. At this time, existing detectors cannot guarantee the reliability of the detection results. Therefore, it is an urgent problem to establish a reasonable and accurate target detection model for the mismatch of fixed interference guidance vectors and design a reliable and effective radar distributed target detector.

[0005] In view of the above analysis, the technical problem that urgently needs to be solved in the prior art is: how to perform reliable and effective target detection in the case of mismatch of the guidance vector of the fixed interference. Summary of the invention

[0006] In view of the problems existing in the prior art, the present invention provides a radar distributed target detection method and system when a fixed interference steering vector is mismatched.

[0007] The present invention is implemented as follows: a radar distributed target detection method when a fixed interference steering vector is mismatched, characterized in that the radar distributed target detection method when a fixed interference steering vector is mismatched comprises:

[0008] S1: Obtain the test data sampled in K units to be detected by the radar detection system with M channels The target signal of interest belongs to the column-full rank matrix The signal subspace of the span is, and the coordinate matrix of the target signal in the signal subspace is The steering vector of the fixed interference has mismatch but belongs to the column full rank matrix Zhang Cheng's interference subspace, the coordinate vector of the fixed interference steering vector in the interference subspace is The conjugate amplitude vector of the fixed interference is Where P is the signal subspace dimension, Q is the interference subspace dimension, and the number of radar detection system channels M can also be called the data model dimension. represents a complex matrix of dimension m×n;

[0009] S2: Obtain auxiliary data sampled by a radar detection system with M channels in L neighboring cells of the cell to be detected

[0010] S3: Using auxiliary data X L Compute the noise covariance matrix Estimated value of

[0011] S4: Under the targetless hypothesis H 0 Under this condition, the probability density function f of the test data X is given according to the estimated value S of the noise covariance matrix R. 0 (X)| R＝S The expression of f 0 (X)| R＝S And the interference subspace matrix J, solve the coordinate vector of the fixed interference steering vector in the data to be tested in the interference subspace spanned by the matrix J Estimated value of Solve the fixed interference conjugate amplitude vector in the measured data Estimated value of Solve the estimated value of the power mismatch γ of the Gaussian colored noise in the measured data and the auxiliary data

[0012] S5: In the case of a target hypothesis H1 Under this condition, the probability density function f of the test data X is given according to the estimated value S of the noise covariance matrix R. 1 (X)| R＝S The expression of f 1 (X)| R＝S , signal subspace matrix H and interference subspace matrix J, solve the coordinate vector of the fixed interference steering vector in the data to be measured in the interference subspace spanned by matrix J Estimated value of Solve the fixed interference conjugate amplitude vector in the measured data Estimated value of Solve the coordinate matrix of the target signal in the signal subspace Estimated value of Solve the estimated value of the power mismatch γ of the Gaussian colored noise in the measured data and the auxiliary data

[0013] S6: Estimated value of the coordinate vector q of the fixed interference steering vector in the test data in the interference subspace spanned by matrix J Estimation of the fixed interference conjugate amplitude vector β in the measured data Estimation of the power mismatch γ between the measured data and the auxiliary data Substitute the probability density function f of the test data X 0 (X)| R＝S In the above example, we get the target-free hypothesis H 0 The maximum likelihood probability density function max of the test data X q，β，γ f 0 (X)| RR＝S

[0014] S7: Estimated value of the coordinate vector q of the fixed interference steering vector in the test data in the interference subspace spanned by the matrix J Estimation of the fixed interference conjugate amplitude vector β in the measured data The estimated value of the coordinate matrix P of the target signal in the signal subspace Estimation of the power mismatch γ between the measured data and the auxiliary data Substitute the probability density function f of the test data X 1 (X)| R＝S In the above example, we get the target hypothesis H 1 The maximum likelihood probability density function max of the test data X q,β,P,γ f 1 (X)| R＝S ;

[0015] S8: According to the no-target hypothesis H 0The maximum likelihood probability density function max of the test data X q,β,γ f 0 (X)| R＝S and the target hypothesis H 1 The maximum likelihood probability density function max of the test data X q,β,P,γ f 1 (X)| R＝S Calculate the detection statistic t based on the generalized likelihood ratio criterion;

[0016] S9: According to the false alarm probability P fa Determine the detection threshold T;

[0017] S10: Compare the detection statistic t based on the generalized likelihood ratio criterion with the detection threshold T to determine whether there is a target signal of interest in the data to be tested X.

[0018] Further, in step S3, the auxiliary data X is used L The estimated value S of the noise covariance matrix R is calculated as: in(·) H represents the conjugate transpose.

[0019] Further, in step S4, the estimated value S of the noise covariance matrix R is given under the targetless hypothesis H 0 The probability density function f of the test data X 0 (X)| R＝S for: where exp(·) represents the natural exponential function, tr(·) represents the matrix trace, |·| represents the absolute value of a scalar or the determinant of a matrix, and X S =S -1 / 2 X, J S =S -1 / 2 J, (·) -1 / 2 Represents matrix square root inversion.

[0020] Further, the estimated value of the coordinate vector q of the fixed interference steering vector in the test data in the interference subspace spanned by the matrix J in step S4 is for: is a matrix The eigenvector corresponding to the maximum eigenvalue of -1 Represents the inversion of a reversible matrix.

[0021] Further, the estimated value of the fixed interference conjugate amplitude vector β in the test data in step S4 is for:

[0022] Further, the estimated value of the power mismatch γ of the Gaussian colored noise in the test data and the auxiliary data in step S4 is for:

[0023] Further, in step S5, when there is a target hypothesis H 1 The probability density function f of the test data X 1 (X)| R＝S for: Among them, H S =S -1 / 2 H.

[0024] Further, the estimated value of the coordinate vector q of the fixed interference steering vector in the test data in the interference subspace spanned by the matrix J in step S5 is for: is a matrix The eigenvector corresponding to the maximum eigenvalue of I M Represents the identity matrix of dimension M.

[0025] Further, the estimated value of the fixed interference conjugate amplitude vector β in the test data in step S5 is for:

[0026] Further, the estimated value of the coordinate matrix P of the target signal in the signal subspace in step S5 is for:

[0027] Further, the estimated value of the power mismatch γ of the Gaussian colored noise in the test data and the auxiliary data in step S5 is for:

[0028] Further, in step S6, there is no target hypothesis H 0 The maximum likelihood probability density function max of the test data X q,β,γ f 0 (X)| R＝S for:

[0029] Further, in step S7, there is a target hypothesis H 1 The maximum likelihood probability density function max of the test data X q,β,P,γ f 1 (X)| R＝S for:

[0030] Further, the detection statistic t based on the generalized likelihood ratio criterion in step S8 is:

[0031] Further, in step S9, according to the false alarm probability P fa The specific steps to determine the detection threshold T are:

[0032] S91: Calculate a detection statistic t based on a generalized likelihood ratio criterion based on test data without a target signal;

[0033] S92: Execute M on step S91 c The Monte Carlo simulation is repeated and the calculated detection statistic t based on the generalized likelihood ratio criterion is stored in the column vector Among them represents a real matrix of dimension m×n;

[0034] S93: Rearrange the elements in vector V in ascending order. The element is the false alarm probability P fa The corresponding detection threshold T, where Indicates rounding down.

[0035] Furthermore, in step S10, the specific process of comparing the size relationship between the detection statistic t based on the generalized likelihood ratio criterion and the detection threshold T to determine whether there is a target signal of interest in the test data X is: if t>T, the target signal of interest exists; if t≤T, the target signal of interest does not exist.

[0036] Another object of the present invention is to provide a radar distributed target detection system when a fixed interference steering vector is mismatched, which implements the radar distributed target detection method when a fixed interference steering vector is mismatched, comprising:

[0037] Test data acquisition module: used to obtain the test data from the unit to be tested;

[0038] Auxiliary data acquisition module: used to obtain auxiliary data from adjacent units;

[0039] No-target signal test data acquisition module: used to obtain the test data from the unit to be detected when there is no target, so as to calculate the detection threshold;

[0040] Parameter acquisition module: used to obtain the signal subspace column full rank matrix, the interference subspace column full rank matrix, and the false alarm probability;

[0041] Covariance matrix estimation module: used to estimate the noise covariance matrix using auxiliary data;

[0042] Parameter estimation module: used to calculate unknown parameters using the estimated noise covariance matrix, the signal subspace column full rank matrix, the interference subspace column full rank matrix and the data to be measured, wherein the unknown parameters include the coordinate vector of the fixed interference steering vector in the interference subspace, the fixed interference amplitude vector, the power mismatch of the Gaussian colored noise in the data to be measured and the auxiliary data, and the coordinate matrix of the target signal in the signal subspace under the assumption of a target;

[0043] Maximum likelihood probability density function calculation module: used to calculate the maximum likelihood probability density function of the test data under the non-target hypothesis and the target hypothesis using the output results of the parameter estimation module;

[0044] A detection statistic calculation module: used to calculate the detection statistic based on the generalized likelihood ratio criterion using the output result of the maximum likelihood probability density function calculation module;

[0045] Detection threshold calculation module: used to calculate the detection threshold based on the no-target signal test data and the false alarm probability;

[0046] Target detection output module: used to compare the detection statistics and the detection threshold, and output the decision result of whether there is a target signal of interest in the data to be tested.

[0047] Another object of the present invention is to provide a computer device, the computer device comprising a memory and a processor, the memory storing a computer program, and when the computer program is executed by the processor, the processor executes the steps of the radar distributed target detection method when the fixed interference steering vector is mismatched.

[0048] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to perform the steps of the radar distributed target detection method when the fixed interference steering vector is mismatched.

[0049] Another object of the present invention is to provide an information data processing terminal, which includes the radar distributed target detection system when the fixed interference steering vector is mismatched.

[0050] In combination with the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solutions to be protected by the present invention are as follows:

[0051] First, the present invention introduces a subspace model to describe the steering vector of the fixed interference. The subspace covers the possible deviations of the steering vector of the fixed interference, thereby ensuring the reliability of the mathematical model of the detection problem.

[0052] The present invention can achieve more reliable and better distributed target detection performance in the case of fixed interference steering vector mismatch. The present invention can still work when the sum of the signal subspace and interference subspace dimensions is close to or equal to the data model dimension, and achieve effective and reliable distributed target detection performance.

[0053] Second, the technical solution of the present invention fills the technical gap in the industry at home and abroad: the existing technical solutions all assume that the fixed interference guidance information is accurately known, and guarantee reliable detection performance when the fixed interference guidance information is assumed to be completely consistent with the real fixed interference guidance information. The present invention fully considers the problems of array calibration error and interference identification error that are common in complex environments, and uses a subspace model to describe the fixed interference guidance vector that may have mismatches, effectively solving the radar distributed target detection problem when the fixed interference guidance vector is mismatched, and eliminating the risk of unreliable target detection results that may be caused by the fixed interference guidance vector.

[0054] The technical solution of the present invention solves a technical problem that people have been eager to solve but have never been successful: the existing technical solutions all have application blind spots: when the sum of the dimensions of the signal subspace and the interference subspace is close to or even equal to the data model dimension, the performance of the existing technical solutions deteriorates sharply or even fails. The present invention uses a subspace model to cover the possible mismatch range of the fixed interference steering vector. The detection statistics of the designed new detector do not include the part that causes the application blind spot of the existing detector. When the sum of the dimensions of the signal subspace and the interference subspace is close to or even equal to the data model dimension, reliable and effective detection performance can still be guaranteed. Therefore, the technical solution of the present invention does not have an application blind spot. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a schematic diagram of the principle of a radar distributed target detection method when a fixed interference steering vector is mismatched provided by an embodiment of the present invention;

[0056] Figure 2 It is a structural framework diagram of a radar distributed target detection system when a fixed interference steering vector is mismatched provided by an embodiment of the present invention;

[0057] Figure 3 is a comparison diagram of detection probabilities of the technical solution provided by the embodiment of the present invention and the prior art solution under different signal-to-noise ratios;

[0058] Figure 4 is a comparison diagram of detection probabilities of the technical solution provided by the embodiment of the present invention and the prior art solution under different interference-to-noise ratios;

[0059] Figure 5 It is a comparison diagram of detection probabilities of the technical solution provided by the embodiment of the present invention and the prior art solution under different interference subspace dimensions. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0061] The radar detection system of the present invention can be applied to multiple industries and fields, especially in scenarios requiring high-resolution detection and anti-interference capabilities. Two specific industrial application embodiments are listed below:

[0062] Example 1: Marine radar detection system

[0063] In the marine environment, radar systems are often used to detect and track marine targets, such as ships, icebergs, and buoys. However, due to the complex fluctuations in the marine environment, changes in weather conditions, and the presence of sea surface reflections and interference sources (for example, sea surface reflection waves, fixed sea surface buildings, communication interference, etc.), radar systems face huge challenges in identifying and tracking marine targets. Traditional radar detectors are susceptible to interference, especially when the position of the interference source changes, which can lead to false alarms and missed alarms. The detection method in the present invention utilizes the deviation of the fixed interference guidance vector covered by the interference subspace, thereby significantly improving the detection performance of the system in an interference environment. The system has higher resolution and anti-interference ability when detecting mobile targets at sea, and is suitable for applications such as coastal patrols and maritime traffic monitoring.

[0064] Example 2: Aerial target monitoring and early warning system

[0065] In the field of airspace monitoring and early warning, radar systems need to monitor multiple targets in the air to deal with potential threats (such as drones, aircraft, missiles, etc.). In practical applications, airborne radar systems are often affected by various fixed interferences (such as interference from fixed buildings and radio transmitters) and environmental noise. Traditional radars have weak detection capabilities in complex environments and are easily affected by interference mismatches, thereby reducing detection accuracy. The system provided by the present invention can reduce the interference effect by using an interference subspace modeling method when there is a mismatch in the fixed interference guidance vector, so that the radar can accurately identify the target signal and improve the detection performance and target recognition rate of the system. The system can be widely used in the fields of airport airspace security monitoring, border air defense monitoring, military early warning systems, etc., and effectively improve the reliability and accuracy of target detection.

[0066] A high-resolution radar detection system with M channels samples test data from K units to be detected. in is the test data in the kth unit to be tested, k = 1, 2, ..., K, represents a complex matrix of dimension m×n.0 In this case, the test data X contains Gaussian colored noise and fixed interference. Gaussian colored noise can be expressed as All its column vectors are independent and identically distributed and have a mean of zero, and the covariance matrix is ​​R t Gaussian distribution. The fixed interference can be expressed as jβ H ,in is the steering vector of the fixed interference, β=[β 1 ,β 2 ,…,β K ] T is the fixed interference conjugate amplitude vector, is the fixed interference amplitude in the kth unit to be detected, (·) T represents transpose, (·) * Denotes conjugate. Due to the existence of radar channel calibration error, interference identification error, etc., the fixed interference steering vector j has mismatch, at this time, a known column full rank matrix can be used The interference subspace of Zhang Cheng is used to cover the possible deviation of the fixed interference steering vector j, that is, the fixed interference steering vector can be remodeled as j = Jq, and then the fixed interference can be described as Jqβ H ,in Characterizes the coordinate vector of the fixed interference steering vector j in the interference subspace. 1 In this case, the test data X also includes the target echo HP, where is the full rank matrix of the known signal subspace, is the coordinate matrix of the target signal, representing the position of the target signal in the signal subspace spanned by the matrix H. In order to estimate the noise covariance matrix R in the test data t , obtain auxiliary data by sampling from L neighboring units of the unit to be detected Auxiliary DataX L Contains only Gaussian colored noise in Represents the auxiliary data in the lth adjacent unit, l = 1, 2, ..., L. Auxiliary data X L All column vectors in are independent and identically distributed and follow a Gaussian distribution with a mean of zero and a covariance matrix of R. Due to the fluctuations in the actual environment, the noise covariance matrix R in the test data t =γQ, where γ represents the noise power mismatch between the test data and the auxiliary data. In summary, the binary hypothesis testing problem of radar distributed targets under the condition of fixed interference guidance vector mismatch can be expressed as:

[0067] The object of the present invention is to improve the detection performance of the detector for radar distributed targets when the fixed interference steering vector is mismatched.

[0068] like Figure 1 As shown, this embodiment provides a radar distributed target detection method when a fixed interference steering vector is mismatched, and the steps of the method include:

[0069] S1: Obtain the test data sampled in K units to be detected by the radar detection system with M channels The target signal of interest belongs to the column-full rank matrix The signal subspace of the span is, and the coordinate matrix of the target signal in the signal subspace is The steering vector of the fixed interference has mismatch but belongs to the column full rank matrix Zhang Cheng's interference subspace, the coordinate vector of the fixed interference steering vector in the interference subspace is The conjugate amplitude vector of the fixed interference is Where P is the signal subspace dimension, Q is the interference subspace dimension, and the number of radar detection system channels M can also be called the data model dimension. represents a complex matrix of dimension m×n;

[0070] S2: Obtain auxiliary data sampled by a radar detection system with M channels in L neighboring cells of the cell to be detected

[0071] S3: Using auxiliary data X L Compute the noise covariance matrix Estimated value of

[0072] S4: Under the targetless hypothesis H 0 Under this condition, the probability density function f of the test data X is given according to the estimated value S of the noise covariance matrix R. 0 (X)| R＝S The expression of f 0 (X)| R＝S And the interference subspace matrix J, solve the coordinate vector of the fixed interference steering vector in the data to be tested in the interference subspace spanned by the matrix J Estimated value of Solve the fixed interference conjugate amplitude vector in the measured data Estimated value of Solve the estimated value of the power mismatch γ of the Gaussian colored noise in the measured data and the auxiliary data

[0073] S5: In the case of a target hypothesis H 1 Under this condition, the probability density function f of the test data X is given according to the estimated value S of the noise covariance matrix R.1 (X)| R＝S The expression of f 1 (X)| R＝S , signal subspace matrix H and interference subspace matrix J, solve the coordinate vector of the fixed interference steering vector in the data to be measured in the interference subspace spanned by matrix J Estimated value of Solve the fixed interference conjugate amplitude vector in the measured data Estimated value of Solve the coordinate matrix of the target signal in the signal subspace Estimated value of Solve the estimated value of the power mismatch γ of the Gaussian colored noise in the measured data and the auxiliary data

[0074] S6: Estimated value of the coordinate vector q of the fixed interference steering vector in the test data in the interference subspace spanned by matrix J Estimation of the fixed interference conjugate amplitude vector β in the measured data Estimation of the power mismatch γ between the measured data and the auxiliary data Substitute the probability density function f of the test data X 0 (X)| R＝S In the above example, we get the target-free hypothesis H 0 The maximum likelihood probability density function max of the test data X q,β,γ f 0 (X)| R＝S ;

[0075] S7: Estimated value of the coordinate vector q of the fixed interference steering vector in the test data in the interference subspace spanned by the matrix J Estimation of the fixed interference conjugate amplitude vector β in the measured data The estimated value of the coordinate matrix P of the target signal in the signal subspace Estimation of the power mismatch γ between the measured data and the auxiliary data Substitute the probability density function f of the test data X 1 (X)| R＝S In the above example, we get the target hypothesis H 1 The maximum likelihood probability density function max of the test data X q,β,P,γ f 1 (X)| R＝S ;

[0076] S8: According to the no-target hypothesis H 0 The maximum likelihood probability density function max of the test data X q,β,γ f 0 (X)| R＝Sand the target hypothesis H 1 The maximum likelihood probability density function max of the test data X q,β,P,γ f 1 (X)| R＝S Calculate the detection statistic t based on the generalized likelihood ratio criterion;

[0077] S9: According to the false alarm probability P fa Determine the detection threshold T;

[0078] S10: Compare the detection statistic t based on the generalized likelihood ratio criterion with the detection threshold T to determine whether there is a target signal of interest in the data to be tested X.

[0079] In step S3, auxiliary data X is used L The estimated value S of the noise covariance matrix R is calculated as: in(·) H represents the conjugate transpose.

[0080] In step S4, the estimated value S of the noise covariance matrix R is given under the targetless hypothesis H 0 The probability density function f of the test data X 0 (X)| R＝S for: where exp(·) represents the natural exponential function, tr(·) represents the matrix trace, |·| represents the absolute value of a scalar or the determinant of a matrix, and X S =S -1*2 X, J S =S -1 / 2 J, (·) -1 / 2 Represents matrix square root inversion.

[0081] The estimated value of the coordinate vector q of the fixed interference steering vector in the test data in the interference subspace spanned by the matrix J in step S4 is for: is a matrix The eigenvector corresponding to the maximum eigenvalue of -1 Represents the inversion of a reversible matrix.

[0082] The estimated value of the fixed interference conjugate amplitude vector β in the test data in step S4 for:

[0083] The estimated value of the power mismatch γ of the Gaussian colored noise in the test data and the auxiliary data in step S4 for:

[0084] In step S5, there is a target hypothesis H 1The probability density function f of the test data X 1 (X)| R＝S for: Among them, H S =S -1 / 2 H.

[0085] The estimated value of the coordinate vector q of the fixed interference steering vector in the test data in the interference subspace spanned by the matrix J in step S5 is for: is a matrix The eigenvector corresponding to the maximum eigenvalue of I M Represents the identity matrix of dimension M.

[0086] The estimated value of the fixed interference conjugate amplitude vector β in the test data in step S5 for:

[0087] The estimated value of the coordinate matrix P of the target signal in the signal subspace in step S5 for:

[0088] The estimated value of the power mismatch γ of the Gaussian colored noise in the test data and the auxiliary data in step S5 for:

[0089] In step S6, there is no target hypothesis H 0 The maximum likelihood probability density function max of the test data X q,β,γ f 0 (X)| R＝S for:

[0090] In step S7, there is a target hypothesis H 1 The maximum likelihood probability density function max of the test data X q,β,P,γ f 1 (X)| R＝S for:

[0091] The detection statistic t based on the generalized likelihood ratio criterion in step S8 is:

[0092] In step S9, according to the false alarm probability P fa The specific steps to determine the detection threshold T are:

[0093] S91: Calculate a detection statistic t based on a generalized likelihood ratio criterion based on test data without a target signal;

[0094] S92: Execute M on step S91 c The Monte Carlo simulation is repeated and the calculated detection statistic t based on the generalized likelihood ratio criterion is stored in the column vector Among them represents a real matrix of dimension m×n;

[0095] S93: Rearrange the elements in vector V in ascending order. The element is the false alarm probability P fa The corresponding detection threshold T, where Indicates rounding down.

[0096] Furthermore, in step S10, the specific process of comparing the size relationship between the detection statistic t based on the generalized likelihood ratio criterion and the detection threshold T to determine whether there is a target signal of interest in the test data X is: if t>T, the target signal of interest exists; if t≤T, the target signal of interest does not exist.

[0097] like Figure 2 As shown, the present invention provides a radar distributed target detection system when a fixed interference steering vector is mismatched, comprising:

[0098] Test data acquisition module: used to obtain the test data from the unit to be tested;

[0099] Auxiliary data acquisition module: used to obtain auxiliary data from adjacent units;

[0100] No-target signal test data acquisition module: used to obtain the test data from the unit to be detected when there is no target, so as to calculate the detection threshold;

[0101] Parameter acquisition module: used to obtain the signal subspace column full rank matrix, the interference subspace column full rank matrix, and the false alarm probability;

[0102] Covariance matrix estimation module: used to estimate the noise covariance matrix using auxiliary data;

[0103] Parameter estimation module: used to calculate unknown parameters using the estimated noise covariance matrix, the signal subspace column full rank matrix, the interference subspace column full rank matrix and the data to be measured, wherein the unknown parameters include the coordinate vector of the fixed interference steering vector in the interference subspace, the fixed interference amplitude vector, the power mismatch of the Gaussian colored noise in the data to be measured and the auxiliary data, and the coordinate matrix of the target signal in the signal subspace under the assumption of a target;

[0104] Maximum likelihood probability density function calculation module: used to calculate the maximum likelihood probability density function of the test data under the non-target hypothesis and the target hypothesis using the output results of the parameter estimation module;

[0105] A detection statistic calculation module: used to calculate the detection statistic based on the generalized likelihood ratio criterion using the output result of the maximum likelihood probability density function calculation module;

[0106] Detection threshold calculation module: used to calculate the detection threshold based on the no-target signal test data and the false alarm probability;

[0107] Target detection output module: used to compare the detection statistics and the detection threshold, and output the decision result of whether there is a target signal of interest in the data to be tested.

[0108] The radar distributed target detection system when the fixed interference steering vector is mismatched provided by the present invention first obtains the test data through the test data acquisition module, obtains the auxiliary data through the auxiliary data acquisition module, and obtains the test data when there is no target through the no-target signal test data acquisition module, which is used for subsequent target detection processing.

[0109] Secondly, the signal subspace column full rank matrix, the interference subspace column full rank matrix and the false alarm probability are obtained from the parameter acquisition module to ensure that the statistical confidence of the detection meets the design requirements to improve the detection accuracy of the system.

[0110] Next, the estimated value of the noise covariance matrix is ​​obtained from the covariance matrix estimation module. The output results of the test data acquisition module, the covariance matrix estimation module and the parameter acquisition module are input into the parameter estimation module, and multiple key parameters are estimated in sequence: the coordinate vector of the fixed interference steering vector in the interference subspace, the fixed interference amplitude vector, the power mismatch of the Gaussian colored noise in the test data and the auxiliary data, and the coordinate matrix of the target signal in the signal subspace under the target assumption.

[0111] Subsequently, the maximum likelihood probability density function calculation module uses the output results of the parameter estimation module and the covariance matrix estimation module to calculate the maximum likelihood probability density function of the test data under the non-target hypothesis and the target hypothesis, respectively, and calculates the detection statistic under the generalized likelihood ratio criterion through the detection statistic calculation module.

[0112] Finally, the detection threshold calculation module calculates the detection threshold based on the test data without target signals and the preset false alarm probability, and compares it with the detection statistic. Based on this comparison result, the target detection output module determines whether there is a target signal of interest in the test data and outputs the final detection decision. This result provides the radar system with reliable distributed target detection capabilities, especially when there is a mismatch in the fixed interference steering vector.

[0113] The present invention can be applied to all target detection fields where part of the fixed interference guidance information is known, such as

[0114] ① Military field: There may be jammers in the area illuminated by the beam of a military detection radar. When the prior information of the jammer is not precisely known, the technical solution of the present invention can be used to model fixed interference, thereby effectively detecting the target.

[0115] ② Civil aviation field: When the airport radar detects the aircraft at the take-off and landing stage and performs tasks such as air traffic control, there may be interference from objects such as fixed buildings within the radar illumination range. The use of the technical solution of the present invention can effectively reduce the impact of interference.

[0116] ③ Traffic field: During the normal movement of the vehicle, when the vehicle-mounted radar illuminates the distant interference target (buildings, trees, etc.), there is a micro-Doppler effect, which causes the vehicle-mounted radar to jitter in locating the interference target, thereby affecting the vehicle-mounted radar's detection of nearby vehicles, pedestrians and other targets of interest. The technical solution of the present invention uses a subspace model to cover the jitter range of the interference target guidance information, which can effectively improve the detection performance of the vehicle-mounted radar for nearby targets.

[0117] ④ Medical field: When detecting abnormal tissues such as tumors and nodules in the human body during medical activities, they may be interfered by the reflected signals from normal tissues. In addition, since the human body and the detection instrument cannot be absolutely still, the guidance information of the interference signal reflected by the normal tissues will jitter within a small range, which is in line with the application scenario of the technical solution of the present invention.

[0118] In summary, the target detection technology involved in the present invention plays an important role in many fields, and the application field of this technology will continue to expand and deepen.

[0119] The detection process of both the prior art and the technology of the present invention involves complex matrix operations, and it is impossible to analyze the advantages and disadvantages of each technology through theoretical analysis. In this case, independent repeated Monte Carlo simulation can be used to simulate the complex detection process and evaluate the detection probability of the prior art and the technology of the present invention under the same environment. Monte Carlo simulation is a numerical calculation method based on probability and statistical theory. In a statistical sense, Monte Carlo simulation is equivalent to theoretical analysis and is a commonly used numerical analysis method in the industry.

[0120] The effect of the present invention on detecting distributed targets when the fixed interference steering vector is mismatched is described below based on a Monte Carlo simulation experiment.

[0121] Figure 3The detection probability comparison diagram of the technical solution of the present invention and the existing one-step generalized likelihood ratio test (OS-GLRT) detector, two-step generalized likelihood ratio test (TS-GLRT) detector, one-step Rao criterion (OS-Rao) detector, two-step Rao criterion (TS-Rao) detector, and one-step Wald criterion (OS-Wald) detector under different signal-to-noise ratios is shown. The simulation conditions are: false alarm probability P fa = 0.001, calculate the number of Monte Carlo simulations for the detection threshold M c =10 5 , the number of radar detection system channels M = 16, the signal subspace dimension P = 2, the interference subspace dimension Q = 8, the amount of data to be measured K = 4, the amount of auxiliary data L = 2M = 32, the signal-to-noise ratio tr[P H H H (γR) -1 HP] changes from 5dB to 30dB in equal intervals, and the interference-to-noise ratio is β H β×q H J H (γR) -1 Jq = 20dB, where γR is the noise covariance matrix in the measured data, the scalar γ is a random positive real number, and the (i, j)th element in the matrix R is ρ |i-j| , 1≤i,j≤M,ρ=0.95. Figure 3 It can be seen that the detection probability of the method proposed in the present invention is higher than that of the existing detector under different signal-to-noise ratios. When both achieve an 80% detection probability, the signal-to-noise ratio required by the method proposed in the present invention is reduced by more than 3dB compared with the existing detector. When the signal-to-noise ratio is large, such as the signal-to-noise ratio ≥ 27dB, the detection probability of the method proposed in the present invention and the existing OS-GLRT and TS-GLRT detectors can reach 100%.

[0122] Figure 4 The comparison chart of detection probability between the technical solution of the present invention and the existing OS-GLRT, TS-GLRT, OS-Rao, TS-Rao, and OS-Wald detectors at different interference-to-noise ratios is shown. The simulation conditions are: false alarm probability P fa = 0.001, calculate the number of Monte Carlo simulations for the detection threshold M c =10 5 , the number of radar detection system channels M = 16, the signal subspace dimension P = 2, the interference subspace dimension Q = 8, the amount of data to be measured K = 4, the amount of auxiliary data L = 2M = 32, and the signal-to-noise ratio is tr[P H H H (γR) -1HP]=20dB, interference-to-noise ratio β H β×q H J H (γR) -1 Jq changes from -30dB to 30dB at equal intervals, where γR is the noise covariance matrix in the measured data, the scalar γ is a random positive real number, and the (i, j)th element in the matrix R is ρ |i-j| , 1≤i,j≤M,ρ=0.95. Figure 4 It can be seen that the detection probability of the existing detector does not change with the change of the interference-to-noise ratio. This is because the existing detectors all project the data to be measured into the orthogonal interference subspace, and therefore have the ability to suppress interference; the method proposed in the present invention does not have the ability to suppress interference, but from the simulation results it can be seen that the detection probability of the method proposed in the present invention fluctuates by no more than 10% with the change of the interference-to-noise ratio, and the detection probability of the method proposed in the present invention under different interference-to-noise ratios is increased by more than 30% on average compared with the existing detector.

[0123] Figure 5 The comparison chart of detection probability between the technical solution of the present invention and the existing OS-GLRT, TS-GLRT, OS-Rao, TS-Rao, and OS-Wald detectors under different interference subspace dimensions is shown. The simulation conditions are: false alarm probability P fa = 0.001, calculate the number of Monte Carlo simulations for the detection threshold M c =10 5 , the number of radar detection system channels M = 16, the signal subspace dimension P = 1, the interference subspace dimension Q increases from 1 to MP = 15 at equal intervals, the amount of data to be measured K = 4, the amount of auxiliary data L = 2M = 32, and the signal-to-noise ratio is tr[P H H H (γR) -1 HP]=20dB, interference-to-noise ratio is β H β×q H J H (γR) -1 Jq = 20dB, where γR is the noise covariance matrix in the measured data, the scalar γ is a random positive real number, and the (i, j)th element in the matrix R is ρ |i-j| , 1≤i,j≤M,ρ=0.95. Figure 5It can be seen that: when the dimension of the interference subspace is small, such as Q≤3, the detection probability of the method proposed in the present invention and the detection probability of the existing OS-GLRT, TS-GLRT, and OS-Wald detectors are basically maintained at close to 100%; when the dimension of the interference subspace continues to increase, the detection probability of all detectors decreases, but the detection probability of the existing detectors rapidly deteriorates to 0%, and the detection probability of the method proposed in the present invention always maintains more than 80%. It is worth noting that when the dimension of the interference subspace Q=MP=15, the signal subspace and the interference subspace constitute the entire observation space. At this time, the existing detectors are all invalid because the data to be tested is projected into the orthogonal interference subspace, but the method proposed in the present invention is still effective and the detection probability is greater than 80%, which proves that the method proposed in the present invention has a wider range of applications.

[0124] It should be noted that the embodiments of the present invention can be implemented by hardware, software, or a combination of software and hardware. The hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. It can be understood by a person of ordinary skill in the art that the above-mentioned devices and methods can be implemented using computer executable instructions and / or contained in a processor control code, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. Such code is provided on the carrier medium. The device and its modules of the present invention can be implemented by hardware circuits such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., can also be implemented by software executed by various types of processors, and can also be implemented by a combination of the above-mentioned hardware circuits and software, such as firmware.

[0125] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with the technical field within the technical scope disclosed by the present invention and within the spirit and principle of the present invention should be covered by the protection scope of the present invention.

Claims

1. A radar distributed target detection method when a fixed interference steering vector is mismatched, characterized in that: The method specifically includes: S1: Obtain the test data sampled in K units to be detected by the radar detection system with M channels The target signal of interest belongs to the column-full rank matrix The signal subspace of the span is, and the coordinate matrix of the target signal in the signal subspace is The steering vector of the fixed interference has mismatch but belongs to the column full rank matrix Zhang Cheng's interference subspace, the coordinate vector of the fixed interference steering vector in the interference subspace is The conjugate amplitude vector of the fixed interference is Where P is the signal subspace dimension, Q is the interference subspace dimension, and the number of radar detection system channels M can also be called the data model dimension. represents a complex matrix of dimension m×n; S2: Obtain auxiliary data sampled by a radar detection system with M channels in L neighboring cells of the cell to be detected S3: Using auxiliary data X L Compute the noise covariance matrix Estimated value of S4: Under the non-target hypothesis H0, the probability density function f0(X) of the test data X is given according to the estimated value S of the noise covariance matrix R | R＝S expression, and according to f0(X)| R＝S And the interference subspace matrix J, solve the coordinate vector of the fixed interference steering vector in the data to be tested in the interference subspace spanned by the matrix J Estimated value of Solve the fixed interference conjugate amplitude vector in the measured data Estimated value of Solve the estimated value of the power mismatch γ of the Gaussian colored noise in the measured data and the auxiliary data S5: Under the target hypothesis H1, the probability density function f1(X) of the test data v is given according to the estimated value S of the noise covariance matrix R| R＝S expression, and according to f1(X)| R＝S , signal subspace matrix H and interference subspace matrix J, solve the coordinate vector of the fixed interference steering vector in the data to be measured in the interference subspace spanned by matrix J Estimated value of Solve the fixed interference conjugate amplitude vector in the measured data Estimated value of Solve the coordinate matrix of the target signal in the signal subspace Estimated value of Solve the estimated value of the power mismatch γ of the Gaussian colored noise in the measured data and the auxiliary data S6: Estimated value of the coordinate vector q of the fixed interference steering vector in the test data in the interference subspace spanned by matrix J Estimation of the fixed interference conjugate amplitude vector β in the measured data Estimation of the power mismatch γ between the measured data and the auxiliary data Substitute the probability density function f0(X) of the test data X| R＝S In the above example, we can obtain the maximum likelihood probability density function max of the test data X under the targetless hypothesis H0. q，β，γ f0(X)| R＝S ; S7: Estimated value of the coordinate vector q of the fixed interference steering vector in the test data in the interference subspace spanned by the matrix J Estimation of the fixed interference conjugate amplitude vector β in the measured data The estimated value of the coordinate matrix P of the target signal in the signal subspace Estimation of the power mismatch γ between the measured data and the auxiliary data Substitute the probability density function f1(X) of the test data X| R＝S In the above example, we can obtain the maximum likelihood probability density function max of the test data X under the target hypothesis H1. q，β，P，γ f1(X)| R＝S ; S8: According to the maximum likelihood probability density function max of the test data X under the targetless hypothesis H0 q，β，γ f0(X)| R＝S And the maximum likelihood probability density function max of the test data X under the target hypothesis H1 q，β，P，γ f1(X)| R＝S Compute the detection statistic based on the generalized likelihood ratio criterion; S9: According to the false alarm probability P fa Determine the detection threshold T; S10: Compare the size relationship between the detection statistic based on the generalized likelihood ratio criterion and the detection threshold T to determine whether there is a target signal of interest in the data to be tested X.

2. The radar distributed target detection method when the fixed interference steering vector is mismatched as claimed in claim 1, characterized in that: In step S3, auxiliary data X is used L , calculate the estimated value S of the noise covariance matrix R as: in(·) H represents the conjugate transpose.

3. The radar distributed target detection method when the fixed interference steering vector is mismatched as claimed in claim 1, characterized in that: In step S4, the probability density function f0(X) of the test data X under the targetless hypothesis H0 is given according to the estimated value S of the noise covariance matrix R. R＝S for: where exp(·) represents the natural exponential function, tr(·) represents the matrix trace, |·| represents the absolute value of a scalar or the determinant of a matrix, and X S =S -1 / 2 X, J S =S -1 / 2 J, (·) -1 / 2 Represents matrix square root inversion; The estimated value of the coordinate vector q of the fixed interference steering vector in the test data in the interference subspace spanned by the matrix J in step S4 is for: is a matrix The eigenvector corresponding to the maximum eigenvalue of -1 Indicates the inversion of a reversible matrix; The estimated value of the fixed interference conjugate amplitude vector β in the test data in step S4 for: The estimated value of the power mismatch γ of the Gaussian colored noise in the test data and the auxiliary data in step S4 for:

4. The radar distributed target detection method when the fixed interference steering vector is mismatched as claimed in claim 1, characterized in that: In the step S5, the probability density function f1(X) of the test data X under the target hypothesis H1| R＝S for: Among them, H S =S -1 / 2 H; The estimated value of the coordinate vector q of the fixed interference steering vector in the test data in the interference subspace spanned by the matrix J in step S5 is for: is a matrix The eigenvector corresponding to the maximum eigenvalue of IM represents the identity matrix of dimension M; The estimated value of the fixed interference conjugate amplitude vector β in the test data in step S5 for: The estimated value of the coordinate matrix P of the target signal in the signal subspace in step S5 for: The estimated value of the power mismatch γ of the Gaussian colored noise in the test data and the auxiliary data in step S5 for:

5. The radar distributed target detection method when the fixed interference steering vector is mismatched as claimed in claim 1, characterized in that: The maximum likelihood probability density function max of the test data X under the targetless hypothesis H0 in step S6 q，β，γ f0(X)| R＝S for:

6. The radar distributed target detection method when the fixed interference steering vector is mismatched as claimed in claim 1, characterized in that: In step S7, there is a maximum likelihood probability density function max of the test data X under the target hypothesis H1. q,β,p,γ f1(X)| R＝S for:

7. The radar distributed target detection method when the fixed interference steering vector is mismatched as claimed in claim 1, characterized in that: The detection statistic based on the generalized likelihood ratio criterion in step S8 is:

8. The radar distributed target detection method when the fixed interference steering vector is mismatched as claimed in claim 1, characterized in that: In step S9, according to the false alarm probability P fa The specific steps to determine the detection threshold T are: S91: Calculate the detection statistic based on the generalized likelihood ratio criterion based on the target-free signal test data; S92: Execute M on step S91 c The Monte Carlo simulation is repeated and the calculated detection statistic based on the generalized likelihood ratio criterion is stored in the column vector Among them represents a real matrix of dimension m ×n; S93: Rearrange the elements in vector V in ascending order. The element is the false alarm probability P fa The corresponding detection threshold T, where Indicates rounding down.

9. The radar distributed target detection method when the fixed interference steering vector is mismatched as claimed in claim 1, characterized in that: In step S10, the specific process of comparing the size relationship between the detection statistic N based on the generalized likelihood ratio criterion and the detection threshold T to determine whether there is a target signal of interest in the test data X is: if N>T, the target signal of interest exists; if N≤T, the target signal of interest does not exist.

10. A radar distributed target detection system based on the fixed interference steering vector mismatch as described in claims 1-9, characterized in that: The system specifically includes: Test data acquisition module: used to obtain the test data from the unit to be tested; Auxiliary data acquisition module: used to obtain auxiliary data from adjacent units; No-target signal test data acquisition module: used to obtain the test data from the unit to be detected when there is no target, so as to calculate the detection threshold; Parameter acquisition module: used to obtain the signal subspace column full rank matrix, interference subspace column full rank matrix, and false alarm probability; Covariance matrix estimation module: used to estimate the noise covariance matrix using auxiliary data; Parameter estimation module: used to calculate unknown parameters using the estimated noise covariance matrix, the signal subspace column full rank matrix, the interference subspace column full rank matrix and the data to be measured, wherein the unknown parameters include the coordinate vector of the fixed interference steering vector in the interference subspace, the fixed interference amplitude vector, the power mismatch of the Gaussian colored noise in the data to be measured and the auxiliary data, and the coordinate matrix of the target signal in the signal subspace under the assumption of a target; Maximum likelihood probability density function calculation module: used to calculate the maximum likelihood probability density function of the test data under the non-target hypothesis and the target hypothesis using the output results of the parameter estimation module; A detection statistic calculation module: used to calculate the detection statistic based on the generalized likelihood ratio criterion using the output result of the maximum likelihood probability density function calculation module; Detection threshold calculation module: used to calculate the detection threshold based on the no-target signal test data and the false alarm probability; Target detection output module: used to compare the detection statistics and the detection threshold, and output the decision result of whether there is a target signal of interest in the data to be tested.

Images