Sensitive subspace signal correction Rao detection method and system under strong clutter

Abstract

The invention provides a sensitive subspace signal correction Rao detection method and system under strong clutter, comprising data acquisition and model construction, parameter estimation, detection statistic construction, target judgment and detection probability of output prediction, and being capable of adaptively compensating the influence of a non-zero mean value. Simulation and actual measurement data verification show that in a strong non-zero mean clutter environment, the detection performance of the method is far better than that of a traditional subspace Rao detector ignoring the mean influence, the performance fluctuation is small, and the robustness is high.

Description

Technical Field

[0001] This invention relates to the field of electronic information technology, and in particular to a method and system for detecting Rao signal correction in sensitive subspace under strong clutter. Background Technology

[0002] In multi-channel adaptive detection, the Rao test, as an asymptotically optimal detection criterion, has been widely applied to various signal detection problems. Compared to the most commonly used generalized likelihood ratio (GLRT) criterion, the Rao criterion only requires estimating the null hypothesis. The Rao criterion has fewer unknowns and therefore lower computational complexity. Furthermore, compared to the GLRT criterion and other criteria, the Rao criterion often exhibits better detection performance for certain problems or under certain parameter settings.

[0003] Traditional subspace Rao detectors perform well against zero-mean Gaussian clutter. However, in some cases, strong clutter exhibits non-zero-mean characteristics, a situation frequently encountered in hyperspectral imaging and real-world scenarios with strong discrete clutter or sidelobe interference. In such cases, the performance of traditional detectors suffers a severe degradation due to model mismatch.

[0004] In existing technologies, research on non-zero mean environments largely focuses on rank-1 signal models. For more general subspace signal models (with dimensions p ≥ 1), effective detection schemes for non-zero mean clutter remain lacking. Furthermore, the inventors discovered that the rigorous Rao test cannot be directly constructed under the aforementioned non-zero mean subspace signal detection problem model due to the singularity of the Fisher information matrix block, which constitutes a fundamental technical obstacle.

[0005] Therefore, there is an urgent need in the field for a novel detector based on the Rao test criterion that can overcome the above-mentioned defects and is specifically suitable for the detection of non-zero mean clutter neutron space signals. This detector should be able to effectively compensate for the influence of non-zero mean and have good engineering feasibility. Summary of the Invention

[0006] The purpose of this invention is to address the shortcomings of the prior art by providing a method and system for Rao detection of sensitive subspace signals under strong clutter, thereby solving the technical problems of traditional subspace Rao detectors failing in non-zero mean environments and the inability to directly apply rigorous Rao testing, and endowing the detector with a keen ability to distinguish signal mismatch.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: This invention provides a method for Rao detection of sensitive subspace signals under strong clutter, comprising the following steps: S1. Data Acquisition and Model Building: Acquiring data containing the signal to be detected. Dimensional data vector to be detected ,as well as indivual Dimensional training sample data vector Based on the data, a complex-domain binary hypothesis testing model is constructed: Null hypothesis : ; Alternative Hypothesis : ; in, This indicates that the data to be detected does not contain the target signal; This indicates that the data to be detected contains the target signal; express A cyclic complex Gaussian distribution; for A non-zero clutter mean vector in dimensionality; for The clutter covariance matrix of dimension; for A known signal matrix of dimension 1; for An unknown signal coordinate vector of dimension; S2, Parameter estimation: Based on the data vector to be detected With training sample data vector Calculate the null hypothesis Lower non-zero clutter mean vector The estimated value With clutter covariance matrix The estimated value ; S3. Construction of detection statistics: Using the non-zero clutter mean vector obtained in S2. The estimated value With clutter covariance matrix The estimated value Construct a detection statistic based on the modified Rao test. ; S4. Target Decision: Based on the preset false alarm probability. Determine the detection threshold The calculated detection statistics With detection threshold Compare; if If, then the target is determined to exist; if If the target does not exist, then it is determined that the target does not exist. S5. For a given effective output signal-to-noise ratio and the aforementioned detection threshold Output the predicted detection probability PD.

[0008] Furthermore, in S2, the non-zero clutter mean vector The estimated value With clutter covariance matrix The estimated value The calculation is as follows: ; ; in, Indicates conjugate transpose; This represents finding the inverse of a matrix.

[0009] Furthermore, in S3, the detection statistic of the modified Rao test is... The expression is: .

[0010] Furthermore, in S4, the detection threshold False alarm probability solved numerically Determined by equations: ; in, Loss factor Under the null hypothesis The probability density function under; For a specific summation function.

[0011] Furthermore, the probability density function and specific summation functions The expression is: ; ; Among them, In use Replace, along with the probability density function Substitute together the false alarm probability In this study, numerical calculation methods were used to determine the false alarm probability preset by the given system. Calculate the detection threshold under the given value .

[0012] Furthermore, in S5, the detection probability PD is: ; in, ; ; ; ; in, To achieve an effective signal-to-noise ratio; The data vector to be detected The signal components in; for factorial; In use Replace, along with Substitute these values ​​into PD, and the predicted detection probability value PD can be obtained through numerical calculation methods. It is an incomplete gamma function.

[0013] Furthermore, a Rao detection system for sensitive subspace signals under strong clutter is implemented using the aforementioned Rao detection method for sensitive subspace signals under strong clutter; it also includes: detection equipment. The detection equipment includes a data interface module for receiving baseband complex signal data from systems such as radar and sonar; One or more processors; Memory, which stores instruction modules executed by the processor, including: Modify the Rao detection statistic calculation module to calculate the detection statistic. ; The target decision module is used to execute the decision logic.

[0014] The beneficial effects of this invention are as follows: In the subspace signal detection problem under non-zero mean clutter in the complex domain, this application successfully derives a modified Rao test statistic, which solves the fundamental problem that the strict Rao test does not exist in this scenario due to the singularity of the Fisher information matrix block, and provides a new and effective solution for this type of problem. The Rao-NMC detector adaptively compensates for the influence of non-zero means by jointly estimating the clutter mean and covariance matrix. Simulation and experimental data verification show that in environments with strong non-zero mean clutter, its detection performance is far superior to that of traditional subspace Rao detectors that ignore the influence of means, exhibiting small performance fluctuations and strong robustness. The detector proposed in this application exhibits high sensitivity to signal mismatch. Its detection probability decreases rapidly when the actual target guidance vector deviates from the preset signal subspace. This characteristic makes it particularly suitable for applications requiring precise differentiation between targets and nearby interference within a specific subspace, effectively reducing the risk of false detection. This application provides an analytical expression (integral form) for the false alarm probability of the detector, which enables the detection threshold to be set accurately and offline according to the preset false alarm probability, ensuring the reliability and predictability of the detection system in practical applications; The detection statistics are entirely composed of basic operations on the sampled data matrix (such as vector dot product and matrix inversion), requiring no iterative optimization or solution. The complex estimation of the signal has low computational complexity. It is easy to implement on existing digital signal processors or field-programmable gate arrays, with low processing latency, meeting real-time processing requirements and high engineering feasibility. Attached Figure Description

[0015] Figure 1 A flowchart of a Rao detection method for sensitive subspace signals under strong clutter; Figure 2 The graph shows the performance comparison of the Rao-NMC detector with the classic subspace signal detectors SGLRT, SRao and SAMF at different signal-to-noise ratios (SCRs). Figure 3 Rao-NMC detector and traditional detector at different angles The following is a comparison chart of detection performance; Figure 4 The detection probability of the detector with respect to the degree of signal mismatch with other detectors Comparison chart of changes. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0017] Please see Figure 1 A method for Rao detection of sensitive subspace signals under strong clutter includes the following steps: S1. Data Acquisition and Model Building: Acquiring data containing the signal to be detected. Dimensional data vector to be detected ,as well as indivual Dimensional training sample data vector Based on the data, a complex-domain binary hypothesis testing model is constructed: Null hypothesis : ; Alternative Hypothesis : ; in, This indicates that the data to be detected does not contain the target signal; This indicates that the data to be detected contains the target signal; express A cyclic complex Gaussian distribution; for A non-zero clutter mean vector in dimensionality; for The clutter covariance matrix of dimension; for A known signal matrix of dimension 1; for An unknown signal coordinate vector of dimension; S2, Parameter estimation: Based on the data vector to be detected With training sample data vector Calculate the null hypothesis Lower non-zero clutter mean vector The estimated value With clutter covariance matrix The estimated value ; S3. Construction of detection statistics: Using the non-zero clutter mean vector obtained in S2. The estimated value With clutter covariance matrix The estimated value Construct a detection statistic based on the modified Rao test. ; S4. Target Decision: Based on the preset false alarm probability. Determine the detection threshold The calculated detection statistics With detection threshold Compare; if If, then the target is determined to exist; if If the target does not exist, then it is determined that the target does not exist. S5. For a given effective output signal-to-noise ratio and the aforementioned detection threshold Output the predicted detection probability PD.

[0018] In S2, the non-zero clutter mean vector The estimated value With clutter covariance matrix The estimated value The calculation is as follows: ; ; in, Indicates conjugate transpose; This represents finding the inverse of a matrix.

[0019] In S3, the detection statistic of the modified Rao test is... The expression is: .

[0020] In S4, the detection threshold False alarm probability solved numerically Determined by equations: ; in, Loss factor Under the null hypothesis The probability density function under; For a specific summation function.

[0021] The probability density function and specific summation functions The expression is: ; ; Among them, In use Replace, along with the probability density function Substitute together the false alarm probability In this study, numerical calculation methods were used to determine the false alarm probability preset by the given system. Calculate the detection threshold under the given value .

[0022] In S5, the detection probability PD is: ; in, ; ; ; ; in, To achieve an effective signal-to-noise ratio; The data vector to be detected The signal components in; for factorial; In use Replace, along with Substitute these values ​​into PD, and the predicted detection probability value PD can be obtained through numerical calculation methods. It is an incomplete gamma function.

[0023] A Rao detection system for sensitive subspace signals under strong clutter is implemented using the aforementioned Rao detection method for sensitive subspace signals under strong clutter; it also includes: detection equipment. The detection equipment includes a data interface module for receiving baseband complex signal data from systems such as radar and sonar; One or more processors; Memory, which stores instruction modules executed by the processor, including: Modify the Rao detection statistic calculation module to calculate the detection statistic. ; The target decision module is used to execute the decision logic.

[0024] Simulation experiment: The performance of the invention was verified using Monte Carlo simulation. In the simulation, the number of array elements was set. Subspace dimension Number of training samples Non-zero mean clutter power in, Defined as .

[0025] Figure 2 The performance comparison between the proposed detector Rao-NMC and the existing classic detector SRao is presented in the absence of signal mismatch. It can be seen that Rao-NMC exhibits the best detection performance in the low signal-to-noise ratio region (SCR < 14 dB).

[0026] In fact, the detection performance of classical detectors is also affected by the square of the cosine of the angle between the signal component and the non-zero mean of the clutter in the whitening space. The impact, which can be seen from Figure 3 It can be seen from this that, with With the increase of [something], the detection probability of the three classic detectors all decreased, but the detection probability of the detector proposed in this invention was not affected. This verifies the effectiveness of the detector proposed in this invention in non-zero mean clutter.

[0027] Figure 3 middle, Represents signal components Non-zero mean of clutter At the angle of the whitened space, when the non-zero mean of the clutter is very close to the signal component, the value is often large, close to 1. This usually occurs when the target azimuth contains strong isolated clutter interference.

[0028] Figure 4The performance comparison of the detector proposed in this invention with two other detectors under signal mismatch is presented: the GLRT-NMC detector designed based on the generalized likelihood ratio (GLRT) and the Wald-NMC detector designed based on the modified Wald criterion. As can be seen from the figure, when there is no signal mismatch (i.e., ... (At that time) The detector proposed in this invention, along with the GLRT-NMC and Wald-NMC detectors, can provide a near 100% detection probability. However, when signal mismatch exists, the detector proposed in this invention exhibits the best mismatch sensitivity, i.e., when the signal mismatch becomes severe ( (Reduced), the detection probability of the detector proposed in this application decreases rapidly, thereby avoiding the detection and processing of signals with large mismatch as signals of interest, and ultimately suppressing signals with large mismatch.

[0029] in, , , , , ; ; Represents signal components With respect to the system's assumed subspace (composed of the signal matrix) The square of the cosine of the angle between the subspaces spanned by each column. The smaller the value, the more severe the signal mismatch.

[0030] The embodiments described above are merely illustrative of implementation methods of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be defined by the appended claims.

Claims

1. A method for detecting Rao (Ray-induced distortion) in a sensitive subspace under strong clutter, characterized in that, Includes the following steps: S1. Data Acquisition and Model Building: Acquiring data containing the signal to be detected. Dimensional data vector to be detected ,as well as indivual Dimensional training sample data vector Based on the data, a complex-domain binary hypothesis testing model is constructed: Null hypothesis : ; Alternative Hypothesis : ; in, This indicates that the data to be detected does not contain the target signal; This indicates that the data to be detected contains the target signal; express A cyclic complex Gaussian distribution; for A non-zero clutter mean vector in dimensionality; for The clutter covariance matrix of dimension; for A known signal matrix of dimension 1; for An unknown signal coordinate vector of dimension; S2, Parameter estimation: Based on the data vector to be detected With training sample data vector Calculate the null hypothesis Lower non-zero clutter mean vector The estimated value With clutter covariance matrix The estimated value ; S3. Construction of detection statistics: Using the non-zero clutter mean vector obtained in S2. The estimated value With clutter covariance matrix The estimated value Construct a detection statistic based on the modified Rao test. ; S4. Target Decision: Based on the preset false alarm probability. Determine the detection threshold The calculated detection statistic With detection threshold Compare; if If, then the target is determined to exist; if If the target does not exist, then it is determined that the target does not exist. S5. For a given effective output signal-to-noise ratio and the aforementioned detection threshold Output the predicted detection probability PD.

2. The Rao detection method for sensitive subspace signals under strong clutter as described in claim 1, characterized in that: In S2, the non-zero clutter mean vector The estimated value With clutter covariance matrix The estimated value The calculation is as follows: ； ； in, Indicates conjugate transpose; This represents finding the inverse of a matrix.

3. The Rao detection method for sensitive subspace signals under strong clutter as described in claim 2, characterized in that: In S3, the detection statistic of the modified Rao test is... The expression is: 。 4. The Rao detection method for sensitive subspace signals under strong clutter as described in claim 3, characterized in that: In S4, the detection threshold False alarm probability solved numerically Determined by equations: ； in, Loss factor Under the null hypothesis The probability density function under; For a specific summation function.

5. The Rao detection method for sensitive subspace signals under strong clutter as described in claim 4, characterized in that: The probability density function and specific summation functions The expression is: ； ； Among them, In use Replace, along with the probability density function Substitute together the false alarm probability In this study, numerical calculation methods were used to determine the false alarm probability preset by the given system. Calculate the detection threshold under the given value .

6. The Rao detection method for sensitive subspace signals under strong clutter as described in claim 5, characterized in that: In S5, the detection probability PD is: ； in, ； ； ； ； in, To achieve an effective signal-to-noise ratio; The data vector to be detected The signal components in; for factorial; In use Replace, along with Substitute these values ​​into PD, and the predicted detection probability value PD can be obtained through numerical calculation methods. It is an incomplete gamma function.

7. A Rao detection system for sensitive subspace signals under strong clutter, characterized in that: The method employs a Rao detection method for sensitive subspace signals under strong clutter as described in any one of claims 1 to 6; it also includes: a detection device. The detection equipment includes a data interface module for receiving baseband complex signal data from systems such as radar and sonar; One or more processors; Memory, which stores instruction modules executed by the processor, including: Modify the Rao detection statistic calculation module to calculate the detection statistic. ; The target decision module is used to execute the decision logic.

Images