Minimum inverse eigenvalue radar signal detection method adopting stochastic resonance preprocessing

Abstract

The invention discloses a minimum inverse eigenvalue radar signal detection method adopting stochastic resonance preprocessing, and the method comprises the steps: carrying out the secondary sampling of a received signal, compressing the frequency of the signal to match a stochastic resonance system, and expanding the adaptability of the stochastic resonance system to a higher-frequency signal. Then performing stochastic resonance preprocessing on the signals, enhancing signal energy under a low signal-to-noise ratio to obtain processed signal information, sending the processed signals to a CFAR signal detector, constructing a covariance matrix of the signals of the reference unit and the detection unit, performing eigenvalue decomposition on the covariance matrix to obtain a minimum inverse eigenvalue, and finally obtaining a minimum inverse eigenvalue; the minimum inverse feature value of the detection unit is compared with the average minimum inverse feature value of the reference unit to determine the existence problem of the signal. According to the method, the problems that a matrix constant false alarm detector based on the maximum characteristic value is easily influenced by noise uncertainty and a constant false alarm detector based on information geometry has relatively high calculation complexity are solved.

Description

Technical Field

[0001] The invention belongs to the field of radar signal detection, and in particular relates to a minimum inverse eigenvalue radar signal detection method using stochastic resonance preprocessing. Background Art

[0002] The constant false alarm detector (CFAR) can effectively maintain a constant false alarm probability in a complex interference environment. In particular, the detection of weak radar signals in a low signal-to-noise ratio environment is an important and arduous task, which is of great significance to both military and civilian fields. Conventional constant false alarm detection methods include the unit average constant false alarm detection algorithm (CA-CFAR) using the original radar signal data, but its false alarm probability is high at the edge of the clutter. In recent years, the constant false alarm detector based on the signal covariance matrix has received widespread attention, namely the matrix constant false alarm detector (Matrix-CFAR). The covariance matrix can process multi-dimensional signals, which makes it highly adaptable when processing complex signal systems. Whether in communications, radar or other signal processing fields, the covariance matrix can provide an effective analysis tool. By analyzing the covariance matrix, we can understand whether there is a linear correlation between signals. The analysis of this linear relationship is crucial for signal detection and helps to identify interactions and potential interference between signals.

[0003] Some matrix constant false alarm detectors based on covariance matrix have been proposed one after another. The constant false alarm detector based on the maximum eigenvalue of the covariance matrix is ​​susceptible to noise uncertainty. The classical information geometry constant false alarm detector has high computational complexity, which limits the detection scenario. In addition, in a low signal-to-noise ratio environment, the signal is easily submerged by noise, resulting in a decrease in the detection efficiency of the above detection method. The stochastic resonance system has been successfully applied to the field of signal detection, but the stochastic resonance system with fixed parameters cannot adapt to complex signal detection scenarios. Therefore, a new technical solution is needed to solve the above problems. Summary of the invention

[0004] In order to overcome the shortcomings of the above-mentioned prior art, the present invention provides a minimum inverse eigenvalue radar signal detection method using stochastic resonance preprocessing, which compresses the frequency of the signal to match the stochastic resonance system by performing secondary sampling on the received signal, thereby expanding the adaptability of the stochastic resonance system to higher frequency signals. Subsequently, the signal is subjected to stochastic resonance preprocessing to enhance the signal energy under low signal-to-noise ratio, obtain the processed signal information, send the processed signal to a constant false alarm (CFAR) signal detector, and construct the covariance matrix of the reference unit and the detection unit signal, perform eigenvalue decomposition on the covariance matrix to obtain the minimum inverse eigenvalue, and compare the minimum inverse eigenvalue of the detection unit with the average minimum inverse eigenvalue of the reference unit to determine the existence of the signal. This method solves the problem that the matrix constant false alarm detector based on the maximum eigenvalue is susceptible to noise uncertainty and the constant false alarm detector based on information geometry has a high computational complexity.

[0005] The technical solution adopted by the present invention is as follows: In a first aspect, the present invention provides a minimum inverse eigenvalue radar signal detection method using stochastic resonance preprocessing, the method comprising: The receiver captures the radar signal and performs secondary sampling on the radar signal to obtain a sampling signal; The sampled signal is subjected to adaptive stochastic resonance preprocessing; The detection unit and each reference unit of the constant false alarm detector are respectively subjected to adaptive stochastic resonance preprocessing, and the covariance matrix of the detection unit and each reference unit is respectively calculated using the average value of the stochastic resonance preprocessing output signal; Perform eigenvalue decomposition on the covariance matrix of each detection unit and reference unit, and calculate the minimum inverse eigenvalue of the detection unit covariance matrix and the covariance matrix of each reference unit; The detection statistic is composed of the minimum inverse eigenvalue of the detection unit covariance matrix and the covariance matrix of each reference unit. ,Will With pre-set threshold For comparison, if Greater than It is determined that the radar signal does not exist. Less than It is determined that a radar signal exists.

[0006] Preferably, the secondary sampling implementation process is: Scale the radar signal and introduce the scale factor , specifically expressed as: Formula (1) in represents the sampled signal after scale transformation, represents the signal amplitude, represents a discrete time series, represents the sampling interval, Indicates the radar signal frequency.

[0007] Preferably, the adaptive stochastic resonance preprocessing implementation process is: The stochastic resonance Langevin equation is formally defined as: Formula (2) In the formula, represents the stochastic resonance preprocessing output signal, Indicates time, is additive Gaussian white noise; is the potential function of the bistable system, and its specific form is as follows: Formula (3) in, and is the potential function parameter.

[0008] Find an approximate solution to the stochastic resonance Langevin equation and obtain the stochastic resonance preprocessing output signal Each sample value in .

[0009] Preferably, the potential function of the bistable system is , , ; Represents the energy of the noise.

[0010] Preferably, the implementation process of performing eigenvalue decomposition on the covariance matrix of each detection unit and reference unit and calculating the minimum inverse eigenvalue of the detection unit covariance matrix and each reference unit covariance matrix is: Note: The covariance matrix of the reference units is , represents the detection unit covariance matrix, express The reference cell covariance matrix; Inverse eigenvalue of the covariance matrix With eigenvalue The relationship satisfies: Formula (4) in , Represents the covariance matrix The dimension of Then the covariance matrix of the reference unit is obtained The relationship between the inverse eigenvalues ​​satisfies: ; The minimum inverse eigenvalue of the covariance matrix of the reference unit is further obtained as: Formula (5) in, and denote the maximum eigenvalue and minimum eigenvalue of the covariance matrix of the reference unit, respectively; Without loss of generality, let the minimum inverse eigenvalue of the detection unit be , the minimum inverse eigenvalues ​​of the reference unit are , all the smallest inverse eigenvalues ​​are independent of each other, obey Distribution, denoted as: , so the joint probability density function of the minimum inverse eigenvalue is expressed as: Formula (6)

[0011] Preferably, the covariance matrix No. The inverse eigenvalues ​​are recorded as: Formula (7) in is the covariance matrix The inverse eigenvalues ​​correspond to the inverse eigenvectors, is the covariance matrix and vector The angle between represents the inner product of vectors, Represents the norm of a vector.

[0012] Preferably, the detection statistic The calculation process is: in is the minimum inverse eigenvalue of the detection unit, is the minimum inverse eigenvalue of the M reference units.

[0013] In a second aspect, the present invention provides a radar signal detection system, comprising: The data acquisition module is responsible for acquiring radar signals and performing secondary sampling on the radar signals to obtain sampling signals; The first calculation module is responsible for performing adaptive stochastic resonance preprocessing on the sampled signal; The second calculation module is responsible for performing adaptive random resonance preprocessing on the detection unit and each reference unit of the constant false alarm detector, and calculating the covariance matrix of the detection unit and each reference unit respectively using the average value of the output signal of the first calculation module; performing eigenvalue decomposition on the covariance matrix of each detection unit and reference unit, and calculating the minimum inverse eigenvalue of the covariance matrix of the detection unit and the covariance matrix of each reference unit; and forming the detection statistic by the minimum inverse eigenvalue of the covariance matrix of the detection unit and the covariance matrix of each reference unit. ; The detection module is responsible for converting the detection statistics With pre-set threshold For comparison, if Greater than It is determined that the radar signal does not exist. Less than It is determined that a radar signal exists.

[0014] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to execute the method described.

[0015] In a fourth aspect, the present invention provides a computing device, comprising a memory and a processor, wherein the memory stores executable code, and when the processor executes the executable code, the described method is implemented.

[0016] Compared with the prior art, the present invention has the following beneficial effects: The present invention rationally utilizes noise through adaptive random resonance, transfers noise energy to the energy of weak signals, and solves the problem of high signal missed detection rate under low signal-to-noise ratio when the traditional algorithm directly performs constant false alarm detection. The present invention constructs the covariance matrix of the preprocessed signal, and uses the minimum inverse eigenvalue of the covariance matrix to replace the original data of the detection unit and the reference unit of the constant false alarm detector, which solves the problem that the performance of the maximum eigenvalue-based algorithm is degraded when facing noise uncertainty and the information geometry algorithm has high complexity, so that the method of the present invention can obtain better signal detection performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of a matrix constant false alarm detector of the present invention.

[0018] Figure 2 It is a schematic diagram of the output of the adaptive stochastic resonance signal of the present invention.

[0019] Figure 3 Schematic diagram of the minimum inverse eigenvalue distribution after stochastic resonance preprocessing of the present invention, wherein (a) is without preprocessing and (b) is after preprocessing.

[0020] Figure 4 This is a schematic diagram of the performance comparison of different algorithms under the conditions of reference unit M=16, number of antennas G set to 20, number of sampling points N=1000, and false alarm probability of 0.001. DETAILED DESCRIPTION

[0021] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0022] This embodiment provides a method for detecting constant false alarm signals using a minimum inverse eigenvalue matrix preprocessed by random resonance. In a low signal-to-noise ratio environment, the noise intensity is much higher than the signal intensity. To address the problem of increased missed detection probability, adaptive random resonance is used to preprocess the signal, transfer the noise energy to the energy of the weak signal, and improve the output signal-to-noise ratio of the system. To address the problem that the maximum eigenvalue is susceptible to noise uncertainty and the geometric calculation complexity of the covariance matrix information is high, a matrix constant false alarm detector with the minimum inverse eigenvalue of the covariance matrix is ​​constructed, which can effectively detect signals while having low calculation complexity. Specifically, the following steps are included: Step S1: The receiver can capture the radar signal through the multi-antenna technology, and perform secondary sampling on the radar signal to obtain a sampled signal; In one implementation manner, the number of antennas of the receiver is , the received radar signal is sampled, and the frequency scaling technology is used to introduce a frequency scaling factor to equivalently reduce the frequency of the signal before the signal is preprocessed by stochastic resonance, thereby expanding the frequency range adapted to the stochastic resonance system, and then the sampled data is sent to the adaptive stochastic resonance system to preprocess the signal.

[0023] In one implementation mode, the secondary sampling implementation process is: Radar signal frequency Perform scale transformation and introduce scale transformation factor , so that the frequency is reduced to the original , the sampled signal after scale transformation It is expressed as: in, represents the signal amplitude, represents a discrete time series, Indicates the sampling interval.

[0024] Step S2, performing adaptive stochastic resonance preprocessing on the sampled signal; In one implementation mode, the adaptive stochastic resonance preprocessing implementation process is: The stochastic resonance system does not require additional noise ,Under the condition of low signal-to-noise ratio, the system utilizes its own noise energy to transfer to the energy of the useful signal, thereby achieving signal enhancement.

[0025] The stochastic resonance Langevin equation is defined as: In the formula, represents the stochastic resonance preprocessing output signal, Indicates time, is the sampling signal obtained in step 1, is additive Gaussian white noise; is the potential function of the bistable system, and its specific form is as follows: in, and is the potential function parameter; Parameter selection of potential function: The potential barrier height of the potential function represents the potential energy value that the system needs to overcome to transition from one stable state to another. The two potential wells of the system potential function are obtained: , and the barrier height: The adaptive stochastic resonance system can set appropriate system parameters according to the current noise power. When the system output has the maximum signal-to-noise ratio, the design criteria of the adaptive stochastic resonance system are: , , ; represents the energy of the noise, It means that the adaptive stochastic resonance system does not need to add additional noise; For the stochastic resonance Langevin equation, the Runge-Kutta algorithm is used to find the approximate solution and obtain the stochastic resonance preprocessed output signal The sample values ​​in .

[0026] Furthermore, the process of using the Runge-Kutta algorithm to find an approximate solution for the stochastic resonance Langevin equation is as follows: in, Preprocessing output signal for stochastic resonance The middle Sub-sampling value, To calculate the step size, Set to the inverse of the system sampling frequency, function , for Middle Sample values, is the average slope, .

[0027] Step S3, performing adaptive stochastic resonance preprocessing on the detection unit and each reference unit of the constant false alarm detector, and calculating the covariance matrix of the detection unit and each reference unit respectively by using the average value of the stochastic resonance preprocessing output signal; CFAR detector detection unit The covariance matrix of , the covariance matrix data of the reference unit . Figure 1 It is a schematic diagram of a matrix constant false alarm detector of the present invention.

[0028] Step S4, performing eigenvalue decomposition on the covariance matrix of each detection unit and reference unit, and calculating the minimum inverse eigenvalue of the detection unit covariance matrix and each reference unit covariance matrix; In one implementation manner, the process of performing eigenvalue decomposition on the covariance matrix of each detection unit and reference unit and calculating the minimum inverse eigenvalue of the detection unit covariance matrix and each reference unit covariance matrix is ​​as follows: Note: The covariance matrix of the reference units is , represents the detection unit covariance matrix, express The reference cell covariance matrix; The angle between the covariance matrix and the vector is , the scalar value with the largest change in direction caused by multiplication of the matrix and the vector is called the inverse eigenvalue. Using this feature, the covariance matrix No. The inverse eigenvalues ​​are recorded as: in is the covariance matrix The inverse eigenvalues ​​correspond to the inverse eigenvectors, Represents the covariance matrix The dimension of is the covariance matrix and vector The angle between represents the inner product of vectors, represents the norm of a vector; The inverse eigenvalue can be solved using the eigenvalue of the covariance matrix. The inverse eigenvalue of the covariance matrix is ​​the same as the eigenvalue The relationship satisfies: Then we get the covariance matrix The relationship between the inverse eigenvalues ​​satisfies: ; The minimum inverse eigenvalue of the covariance matrix is ​​further obtained as: in, and Respectively represent the maximum and minimum eigenvalues ​​of the covariance matrix; Without loss of generality, let the minimum inverse eigenvalue of the detection unit be , the minimum inverse eigenvalues ​​of the reference unit are , all the smallest inverse eigenvalues ​​are independent of each other, obey Distribution, denoted as: , so the joint probability density function of the minimum inverse eigenvalue is expressed as:

[0029] Step S5: a detection statistic consisting of the minimum inverse eigenvalue of the detection unit covariance matrix and each reference unit covariance matrix ,Will With pre-set threshold For comparison, if Greater than Then it is determined that the radar signal does not exist, recorded as ,like Less than Then it is determined that the radar signal exists, recorded as .

[0030] In one implementation mode, the detection statistic The calculation process is: in is the minimum inverse eigenvalue of the detection unit, is the minimum inverse eigenvalue of the M reference units.

[0031] In one implementation manner, the threshold The value is determined according to the distribution of the minimum inverse eigenvalue, which is generally less than 1. The value range determines the false alarm probability of the constant false alarm detector. .

[0032] The detection statistic constructed by the present invention The probability density function of is independent of the noise parameters and can maintain the constant false alarm probability property. The following joint probability density function is obtained by using the Jacobian transformation of the multidimensional conditional probability density function: make is the inverse function of T, then: Therefore, the Jacobian matrix is: In the formula, ,therefore is an upper triangular matrix with the determinant , so the joint probability density function can be further expressed as: Therefore, the test statistic The probability density function of As shown below: Then we get the false alarm probability for: It can be seen from the above formula that the false alarm probability of the detection statistic proposed by the present invention is independent of the noise parameter and can maintain a constant false alarm probability property.

[0033] This embodiment is also compared with the following constant false alarm detection methods: a traditional Fourier transform constant false alarm detector (FFT-CFAR), a maximum eigenvalue matrix constant false alarm detector (ME-MD), a maximum and minimum eigenvalue ratio matrix constant false alarm detector (MME-MD), a Log-Euclidean mean detector (Log-Euclidean Mean), and a minimum inverse eigenvalue matrix constant false alarm detector (MA-MD) without preprocessing.

[0034] Figure 2 Under the condition of -10dB, the signal to be detected is submerged in the background noise, and the periodic fluctuation of the signal cannot be observed in the time domain. After the random resonance processing output, the periodicity of the signal is clearly observed. As the noise energy is transferred to the signal energy, its amplitude changes. and The results of the minimum inverse eigenvalue difference before and after preprocessing under the condition are as follows Figure 3 Middle (a) - Figure 3 As shown in (b), the minimum inverse eigenvalue difference after stochastic resonance preprocessing used in the present invention is larger than the direct detection difference, making it easier to distinguish and conditions, increasing the probability of detection.

[0035] Figure 4The performances of different algorithms are compared under the conditions that the reference unit M=16, the number of antennas G is set to 20, and the number of sampling points N=1000. The detection effect of the present invention under low signal-to-noise ratio is better than that of other detection methods.

[0036] In summary, from the analysis of the simulation effect diagram, it can be seen that the minimum inverse eigenvalue constant false alarm detection method based on adaptive bistable stochastic resonance proposed in the present invention improves the signal detection probability under low signal-to-noise ratio.

[0037] This embodiment also provides a radar signal detection system, including: The data acquisition module is responsible for acquiring radar signals and performing secondary sampling on the radar signals to obtain sampling signals; The first calculation module is responsible for performing adaptive stochastic resonance preprocessing on the sampled signal; The second calculation module is responsible for performing adaptive random resonance preprocessing on the detection unit and each reference unit of the constant false alarm detector, and calculating the covariance matrix of the detection unit and each reference unit respectively using the average value of the output signal of the first calculation module; performing eigenvalue decomposition on the covariance matrix of each detection unit and reference unit, and calculating the minimum inverse eigenvalue of the covariance matrix of the detection unit and the covariance matrix of each reference unit; and forming the detection statistic by the minimum inverse eigenvalue of the covariance matrix of the detection unit and the covariance matrix of each reference unit. ; The detection module is responsible for converting the detection statistics With pre-set threshold For comparison, if Greater than It is determined that the radar signal does not exist. Less than It is determined that a radar signal exists.

[0038] An embodiment of the present invention provides an electronic device. Specifically, the electronic device includes a memory and a processor. The memory stores executable code. When the processor executes the executable code, the method described in any one of the implementation modes is implemented.

[0039] The memory may include high-speed random access memory (RAM), and may also include non-volatile memory (Non-volatile Memory), such as at least one disk storage. The communication connection between the system network element and at least one other network element is realized through at least one communication interface (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used.

[0040] The bus may be an ISA bus, a PCI bus or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc.

[0041] Among them, the memory is used to store the program, and the processor executes the program after receiving the execution instruction. The method executed by the device for flow process definition disclosed in any embodiment of the above-mentioned embodiment of the present invention can be applied to the processor or implemented by the processor.

[0042] The processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in the processor or an instruction in the form of software. The above processor may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present invention can be implemented or executed. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in conjunction with the embodiments of the present invention can be directly embodied as being executed by a hardware decoding processor, or executed by a combination of hardware and software modules in a decoding processor. The software module may be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0043] The computer program product of the readable storage medium provided in the embodiment of the present invention includes a computer-readable storage medium storing program code, and the instructions included in the program code can be used to execute the methods described in the previous method embodiments. The specific implementation can be referred to the previous method embodiments, which will not be repeated here.

[0044] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.

[0045] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed by the present invention, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A minimum inverse eigenvalue radar signal detection method using stochastic resonance preprocessing, characterized in that: The method comprises: The receiver captures the radar signal and performs secondary sampling on the radar signal to obtain a sampling signal; The sampled signal is subjected to adaptive stochastic resonance preprocessing; The detection unit and each reference unit of the constant false alarm detector are respectively subjected to adaptive stochastic resonance preprocessing, and the covariance matrix of the detection unit and each reference unit is respectively calculated using the average value of the stochastic resonance preprocessing output signal; Perform eigenvalue decomposition on the covariance matrix of each detection unit and reference unit, and calculate the minimum inverse eigenvalue of the detection unit covariance matrix and the covariance matrix of each reference unit; The detection statistic is composed of the minimum inverse eigenvalue of the detection unit covariance matrix and the covariance matrix of each reference unit. ,Will With pre-set threshold For comparison, if Greater than It is determined that the radar signal does not exist. Less than It is determined that a radar signal exists.

2. The method according to claim 1, characterized in that: The secondary sampling implementation process is: Scale the radar signal and introduce the scale factor , specifically expressed as: Formula (1) in represents the sampled signal after scale transformation, represents the signal amplitude, represents a discrete time series, represents the sampling interval, Indicates the radar signal frequency.

3. The method according to claim 1, characterized in that: The adaptive stochastic resonance preprocessing implementation process is: The stochastic resonance Langevin equation is formally defined as: Formula (2) In the formula, represents the stochastic resonance preprocessing output signal, Indicates time, is additive Gaussian white noise; is the potential function of the bistable system, and its specific form is as follows: Formula (3) in, and is the potential function parameter; Find an approximate solution to the stochastic resonance Langevin equation and obtain the stochastic resonance preprocessing output signal Each sample value in .

4. The method according to claim 3, characterized in that: The potential function of the bistable system , , ; Represents the energy of the noise.

5. The method according to claim 1, characterized in that: The implementation process of performing eigenvalue decomposition on the covariance matrix of each detection unit and reference unit and calculating the minimum inverse eigenvalue of the detection unit covariance matrix and each reference unit covariance matrix is: Note: The covariance matrix of the reference units is , represents the detection unit covariance matrix, express The reference cell covariance matrix; Inverse eigenvalue of the covariance matrix With eigenvalue The relationship satisfies: Formula (4) in , Represents the covariance matrix The dimension of Then the covariance matrix of the reference unit is obtained The relationship between the inverse eigenvalues ​​satisfies: ; The minimum inverse eigenvalue of the covariance matrix of the reference unit is further obtained as: Formula (5) in, and denote the maximum eigenvalue and minimum eigenvalue of the covariance matrix of the reference unit, respectively; Without loss of generality, let the minimum inverse eigenvalue of the detection unit be , the minimum inverse eigenvalues ​​of the reference unit are , all the smallest inverse eigenvalues ​​are independent of each other, obey Distribution, recorded as: , so the joint probability density function of the minimum inverse eigenvalue is expressed as: Formula (6).

6. The method according to claim 5, characterized in that: The covariance matrix No. The inverse eigenvalues ​​are recorded as: Formula (7) in is the covariance matrix The inverse eigenvalues ​​correspond to the inverse eigenvectors, is the covariance matrix and vector The angle between represents the inner product of vectors, Represents the norm of a vector.

7. The method according to claim 1, characterized in that: The detection statistic The calculation process is: in is the minimum inverse eigenvalue of the detection unit, is the minimum inverse eigenvalue of the M reference units.

8. A radar signal detection system for implementing the method according to any one of claims 1 to 7, characterized in that: include: The data acquisition module is responsible for acquiring radar signals and performing secondary sampling on the radar signals to obtain sampling signals; The first calculation module is responsible for performing adaptive stochastic resonance preprocessing on the sampled signal; The second calculation module is responsible for performing adaptive stochastic resonance preprocessing on the detection unit and each reference unit of the constant false alarm detector, and calculating the covariance matrix of the detection unit and each reference unit respectively using the average value of the output signal of the first calculation module; Perform eigenvalue decomposition on the covariance matrix of each detection unit and reference unit, and calculate the minimum inverse eigenvalue of the detection unit covariance matrix and the covariance matrix of each reference unit; The detection statistic is formed by the minimum inverse eigenvalue of the detection unit covariance matrix and the covariance matrix of each reference unit ; The detection module is responsible for converting the detection statistics With pre-set threshold For comparison, if Greater than It is determined that the radar signal does not exist. Less than It is determined that a radar signal exists.

9. A computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to execute the method according to any one of claims 1 to 7.

10. A computing device, comprising a memory and a processor, wherein the memory stores executable code, and when the processor executes the executable code, the method according to any one of claims 1 to 7 is implemented.

Images