A DOA estimation method based on sea surface scattering area multipath equivalence

Abstract

The application provides a DOA estimation method based on a sea surface scattering area multipath equivalence, constructs a multipath signal equivalence model based on the spatial coherence of the multipath signal, respectively equivalences multiple multipath signals from a sea surface diffuse reflection area and a mirror surface reflection area, thereby converts a DOA estimation problem into a parameter optimization problem, and finally solves the optimization problem by using a genetic algorithm to obtain target angle information. The experimental simulation results show that, compared with a traditional method, the multipath angle measurement model constructed by the application is more suitable for a sea surface environment, and the algorithm based on the constructed multipath model has higher resolution and estimation accuracy than other traditional DOA estimation algorithms, and can better meet the angle measurement requirements of the sea surface environment.

Description

Technical Field

[0001] This invention relates to the field of radar array signal processing, specifically to a DOA estimation method based on the multipath equivalence of the sea surface scattering region. Background Technology

[0002] DOA estimation for low-elevation targets at sea has always been a research hotspot and challenge in the field of radar array signal processing, with wide applications across various sectors. For naval vessels, radar is the "eyes that can see a thousand miles," playing a crucial role in modern high-tech warfare. However, unlike ordinary shore-based radar, the sea environment is complex and variable, making target echo signals more susceptible to modulation from the sea surface. When shipborne radar performs DOA estimation for low-elevation targets, a portion of the beam illuminates the sea surface, resulting in complex multipath effects. This means that in the sea environment, in addition to receiving the target's direct wave signal and its specular reflection signal, the radar also receives multipath signals from the diffuse reflection region of the sea surface. Therefore, the multipath signal model in the sea environment no longer matches the traditional specular reflection model, causing traditional DOA estimation algorithms to fail.

[0003] For angle measurement in marine environments, most algorithms ignore multipath signals in diffuse reflection regions, approximating the angle measurement model as a specular reflection model. Within acceptable error ranges, traditional angle measurement algorithms, including but not limited to typical MUSIC, SSMUSIC, and compressed sensing algorithms, are used for DOA estimation. These algorithms perform relatively well in low sea state conditions, but their performance deteriorates sharply with increasingly complex sea states. Therefore, establishing a multipath model that considers the diffuse reflection region of the sea surface and researching new low-elevation angle altimeter algorithms based on this model can effectively improve the estimation performance of low-elevation angle target signals in real-world marine environments. Summary of the Invention

[0004] The technical problem this invention aims to solve is the mismatch in traditional radar angle measurement models due to complex multipath effects in complex sea surface environments. It provides a DOA estimation method based on the multipath equivalence of the sea surface scattering region. This method utilizes the spatial coherence of neighboring multipath signals to propose a multipath equivalence model, performing equivalent processing on multiple multipath signals from the specular and diffuse reflection regions of the sea surface. The traditional subspace-based DOA estimation problem is transformed into a multi-parameter optimization problem, with reasonable constraints on the parameters to be solved, followed by optimization using a genetic algorithm. Experimental simulation results show that, in a sea surface environment, compared with traditional angle measurement methods, this patented method achieves more accurate estimation results for low-elevation targets under conditions of low signal-to-noise ratio and low snapshot number, demonstrating good engineering application value.

[0005] To address the shortcomings of existing technologies, the present invention aims to achieve the above objectives. The technical solution adopted by the present invention is as follows:

[0006] A DOA estimation method based on the multipath equivalence of the sea surface scattering region is presented in detail below:

[0007] Step 1), set the radar array to sample the received signal:

[0008] The radar employs a uniform linear array composed of M elements with an element spacing of md, where m = 0, 1, 2, ..., M-1, and d is half the wavelength of the incident signal. Considering only the case of a single target arrival, the signal data vector received by the radar at time t is: ,in This indicates the direct wave component of the received signal. This represents the received multipath signal components, where P1 and P2 represent the number of incoming multipath signals in the specular reflection region and the number of incoming multipath signals in the diffuse reflection region, respectively. In the formula Represents the transpose of a matrix. , and These represent the direct wave angle steering vector, the specular reflection angle steering vector, and the diffuse reflection angle steering vector, respectively. and Let be the reflection coefficients of the i-th multipath signal from the specular reflection region and the j-th multipath signal from the diffuse reflection region, respectively. Let n(t) be the real envelope of the narrowband signal source incident on the uniform linear array of the radar, and let n(t) represent the additive white Gaussian noise vector received by the array.

[0009] Step 2): Based on the spatial coherence of adjacent multipath signals, the multipath signals from the specular reflection region and the multipath signals from the diffuse reflection region are processed equivalently to construct an equivalent multipath signal model of the sea surface.

[0010] Step 3) Verify the effectiveness of the equivalent multipath model using compressed sensing and least squares methods.

[0011] Step 4) Based on the sea surface multipath equivalent model, apply reasonable constraints to the parameters to be determined.

[0012] Step 5) Based on the sea surface equivalent multipath model and the constraints obtained in Step 4), the DOA estimation problem for low elevation angle targets on the sea surface is transformed into a conditionally constrained multi-parameter optimization problem.

[0013] Step 6) Use a genetic algorithm to optimize and solve the multi-parameter optimization problem with constraints: convert the cost function into a fitness function and initialize the population with binary encoding; convert the binary encoding of the parameters to be solved into decimal within the feasible region and calculate the fitness value of each individual in the population; perform a series of genetic operations; introduce the excellent individuals from the original population into the next generation of the newly generated population to update the population; repeat the iteration until the termination condition is met and output the optimal result.

[0014] Furthermore, the specific steps of step 2) are as follows:

[0015] Step 2.1), the raw signal received by the radar Normalization is performed to obtain the normalized signal vector:

[0016]

[0017]

[0018] ,

[0019] in, This represents the maximum value among the vector elements after taking the modulo.

[0020] Step 2.2), based on the idea of ​​multipath vector synthesis, using... and Replace the normalized multipath components of the specular reflection region and the diffuse reflection multipath components in step 2.1) respectively. Then construct an equivalent multipath signal model: ,in, Normalized noise that includes equivalent fitting error; It is the virtual equivalent reflection coefficient of the direct waveguide vector; and These are the complex equivalent reflection coefficients for the specular reflection region and the diffuse reflection region, respectively;

[0021] Furthermore, the specific steps of step 3) are as follows:

[0022] Step 3.1), Set the spatial sampling angle A complete dictionary matrix has been constructed. ;

[0023] Step 3.2), construct the optimization function based on the overcomplete dictionary matrix:

[0024]

[0025] in The vector sum of multiple multipath signals. Let L0 be the norm of the vector. Let be a sparse vector to be estimated containing only one non-zero element;

[0026] Step 3.3) involves combining each column vector in the overcomplete matrix with the vector sum of the multipath signals. Performing least squares calculations, we obtain the following formula: ,i=1,...,T. Where represent The only non-zero element in the equation (equivalent multipath coefficient);

[0027] Step 3.4) Substitute the results obtained from step 3.3) into step 3.2) to find the sampling angle corresponding to the minimum value, which is the optimal equivalent multipath angle.

[0028] Furthermore, the specific steps of step 4) are as follows:

[0029] Step 4.1): Compare and match the normalized expression from Step 2.1) with the equivalent multipath signal model from Step 2.2) to obtain:

[0030] ;

[0031] Step 4.2): Using the trigonometric Cauchy inequality, simplify the expression in step 4.1) to obtain:

[0032] ,

[0033] The amplitude range of each equivalent multipath reflection coefficient can be limited to Inside, the phase is set Within the range;

[0034] Step 4.3): Since DOA estimation occurs after target detection, the incoming target signal only exists within half the radar's 3dB beamwidth. Therefore, the direct wave angle is limited to... .in This represents the radar's 3dB beamwidth.

[0035] Step 4.4): Since the multipath region of specular reflection is located near the specular reflection angle, the source from the mirror can also be set as... ;

[0036] Step 4.5): Since the multipath region of specular reflection is located near the specular reflection angle, and the optimal equivalent specular reflection angle is located within the multipath region, the equivalent specular reflection angle can be set to... ;

[0037] Step 4.6): The diffuse multipath region is usually larger than the specular multipath region, but multipath echo signals exceeding half the width of the radar's first null beamwidth can be ignored, and the optimal equivalent diffuse reflection angle is also within the diffuse multipath region. Therefore, the equivalent specular reflection angle can be set to... ,in This represents the first null beamwidth of the radar.

[0038] Furthermore, the specific steps of step 5) are as follows:

[0039] Step 5.1), combined with steps 2) and 4), yields the following constrained multi-parameter optimization problem:

[0040] .

[0041] Furthermore, the detailed steps of step 6 are as follows:

[0042] Step 6.1), therefore, according to the genetic algorithm, the fitness function is constructed as follows:

[0043] ,

[0044] Each individual in the population corresponds to 8 search parameters. Each individual in the population is represented as follows: ,in Let N be the i-th individual in the population, and N be the number of individuals in the population. These are the straight wave angle, the mirror equivalent multipath angle, and the diffuse equivalent multipath angle, respectively. This represents the amplitude and phase of the specular equivalent reflection coefficient. This represents the amplitude and phase of the equivalent reflection coefficient. The equivalent reflection coefficient of the direct wave;

[0045] Step 6.2) Using binary encoding, each parameter in the individual is sequentially concatenated with 16-bit binary codes to form a decoded individual;

[0046] Step 6.3): For each individual in the population, perform decimal decoding within the constraints, calculate the fitness value, and normalize it.

[0047] Step 6.4): Based on the fitness value of each individual in the population, after selecting the best individual in the current population, perform roulette wheel selection, uniform crossover, and mutation operations on the individuals in the original population in turn.

[0048] Step 6.5) The best individual from the previous generation is retained and replaced with the worst individual in the current generation. Then, the genetic iteration is carried out continuously so that more excellent individuals are retained.

[0049] Step 6.6): Set the maximum number of iterations. When the termination condition is met, decode the current best individual according to the prescribed encoding method and output the optimal target angle estimate under the current snapshot.

[0050] Step 6.6) Repeat steps 6.2) to 6.6) to calculate the average optimal target angle estimate from multiple snapshots.

[0051] Compared with the prior art, the present invention adopts the above technical solution and has the following advantages:

[0052] The equivalent multipath model of the scattering region established in this invention is more suitable for complex sea surface backgrounds than the traditional low elevation angle measurement model that only considers specular reflection. From the perspective of the model, it improves the adaptability to complex sea surfaces. Under the conditions of low signal-to-noise ratio and low snapshot number, it has higher accuracy than traditional angle measurement algorithms, thereby improving the angle measurement performance in the sea surface environment. Attached Figure Description

[0053] Figure 1 This is a schematic diagram of a real multipath angle measurement model of radar in a sea surface environment;

[0054] Figure 2 This is a scatter plot of target angle estimation using the DOA estimation method described in this invention under different Monte Carlo experiments;

[0055] Figure 3 This is a graph showing the relationship between the relative error between the equivalent multipath model constructed in this invention and the actual multipath signal and the equivalent angle.

[0056] Figure 4 This is a performance comparison chart of the DOA estimation method described in this invention and the traditional DOA estimation method under different signal-to-noise ratios. Detailed Implementation

[0057] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0058] The DOA estimation method based on multipath equivalence in the sea surface scattering region described in this invention includes the following process:

[0059] First embodiment: as Figure 3 As shown, a scatter plot of target angle estimation using this method in 100 Monte Carlo experiments is presented. First, we constructed a uniform linear array of 15 radar elements with an element spacing of 0.5 wavelengths per meter. The radar's 3dB beamwidth is also shown. First zero-point beamwidth The signal-to-noise ratio is 10, and the number of snapshots is 20. The spatial sampling angle range is [-10, 0], and the number of snapshots is 100. The target arrival angle is 1.81°. The number of arrivals in both the specular reflection region and the diffuse reflection region is 3, respectively: -5.5°, -4.27°, -3.2° and -3.1°, -1.51°, -1.21°. [Perturbation coefficient...] .

[0060] Based on the above conditions, the DOA of the target is estimated using the method proposed in this patent. The specific implementation process is as follows:

[0061] Step 1), set the radar array to sample the received signal:

[0062] The signal data vector received by the radar at time t is: , t=1,...,20 where This indicates the direct wave component of the received signal. This represents the received multipath signal component. In the formula... Represents the transpose of a matrix. , and These represent the direct wave angle steering vector, the specular reflection angle steering vector, and the diffuse reflection angle steering vector, respectively. and Let be the reflection coefficients of the i-th multipath signal from the specular reflection region and the j-th multipath signal from the diffuse reflection region, respectively. Let n(t) be the real envelope of the narrowband signal source incident on the uniform linear array of the radar, and let n(t) represent the additive white Gaussian noise vector received by the array.

[0063] Step 2), the raw signal received by the radar Normalization is performed to obtain the normalized signal vector, specifically:

[0064]

[0065]

[0066] ,

[0067] in, This represents the maximum value among the vector elements after taking the modulo.

[0068] Step 3), based on the idea of ​​multipath vector synthesis, using... and The normalized multipath components of the multiple specular reflection regions and the multiple diffuse reflection multipath components from step 2) are replaced respectively. An equivalent multipath signal model is then constructed. ,in Normalized noise that includes equivalent fitting error; It is the virtual equivalent reflection coefficient of the direct waveguide vector; and These are the complex equivalent reflection coefficients for the specular reflection region and the diffuse reflection region, respectively;

[0069] Step 4), Set the spatial sampling angle A complete dictionary matrix has been constructed. ;

[0070] Step 5), construct the optimization function based on the overcomplete dictionary matrix:

[0071] ,

[0072] in The vector sum of multiple multipath signals. Let L0 be the norm of the vector. Let be a sparse vector to be estimated containing only one non-zero element;

[0073] Step 6) Connect each column vector in the overcomplete matrix to the vector sum of the multipath signals. Performing least squares calculations, we obtain the following formula: ,i=1,...,100. Where represent The only non-zero element in the equation (equivalent multipath coefficient);

[0074] Step 7), substitute the results obtained in step 5) into the sampling angle corresponding to the minimum value obtained in step 6), which is the optimal equivalent multipath angle.

[0075] Second embodiment: as Figure 2 As shown, the relative error between the proposed equivalent multipath model and the actual multipath model as the equivalent multipath angle changes was investigated. Due to the presence of the sea surface, different amplitudes and phases of different multipath signals are modulated; therefore, the reflection coefficient of the multipath signal in the simulation is expressed by the following formula:

[0076] , ,

[0077] in To ensure uniform distribution, This represents the disturbance coefficient.

[0078] To verify the effectiveness of the equivalent multipath model, we introduce relative error as a criterion for evaluating it. The relative error is defined as follows:

[0079] , ,

[0080] Figure 2 The simulation parameters were set as follows: 15 radar array elements, a spacing d of 1 meter (half the signal wavelength), 3 incoming multipath propagation paths in the specular reflection region with incoming angles of -3.1°, -1.51°, and -1.21° respectively. The same 3-fold multipath propagation path propagation path was used in the diffuse reflection region with incoming angles of -5.5°, -4.27°, and -3.2° respectively. The number of spatial sampling angles was set to 100, within the range [-10, 0]. The disturbance coefficient was also set. =3; The specific steps can be completed through steps 4) to 7). As can be seen from the figure, the relative errors of the specular multipath equivalent and the diffuse multipath equivalent are both much less than 0.1, which verifies that the multipath equivalent model is reasonable. Furthermore, it was found that the optimal multipath equivalent arrival angle is always between the echo angles of the specular reflection signal and the diffuse reflection signal;

[0081] Step 8): Compare and match the normalized expression from Step 2) with the equivalent multipath signal model from Step 3). The result is:

[0082] ,

[0083] Step 9), using the trigonometric Cauchy inequality, simplify the expression in step 8) to obtain:

[0084] ,

[0085] The amplitude range of each equivalent multipath reflection coefficient can be limited to Inside, the phase is set Within the range;

[0086] Step 10): Since DOA estimation occurs after target detection, the incoming target signal only exists within half the radar's 3dB beamwidth. Therefore, the direct wave angle is limited to... .in This represents the radar's 3dB beamwidth.

[0087] Step 11): Since the multipath region of mirror reflection is located near the mirror reflection angle, and the optimal equivalent mirror reflection angle is located within the multipath region, the equivalent mirror reflection angle can be set to... ;

[0088] Step 12): The diffuse multipath region is usually larger than the specular multipath region, but multipath echo signals exceeding half the width of the radar's first null beamwidth can be ignored, and the optimal equivalent diffuse reflection angle is also within the diffuse multipath region. Therefore, the equivalent specular reflection angle can be set to... ,in This is the first null beamwidth of the radar;

[0089] Step 13), combining steps 3) and 8) to 12), yields the following constrained multi-parameter optimization problem:

[0090] ;

[0091] Step 14): Using the cost function from Step 13, construct the fitness function for the genetic algorithm.

[0092] This patent searches for a maximum value, therefore the individual with the highest fitness in the population is considered the best individual in the current population.

[0093] Each individual in the population corresponds to 8 search parameters. Therefore, each individual in the population is represented as: ;in Let N be the i-th individual in the population, and N be the number of individuals in the population. These are the straight wave angle, the mirror equivalent multipath angle, and the diffuse equivalent multipath angle, respectively. This represents the amplitude and phase of the specular equivalent reflection coefficient. This represents the amplitude and phase of the equivalent reflection coefficient. The equivalent reflection coefficient of the direct wave;

[0094] Step 15): In the genetic algorithm, the encoding operation represents the feasible solution to the problem as an individual structure that can be evaluated by the fitness function. The patent uses binary encoding, representing each individual in the population as a binary string consisting of 0s and 1s. In Example 1, to ensure search accuracy, the eight parameters are represented using 8×16 bits of binary data.

[0095] Step 16): For each individual in the population, perform decimal decoding within the constraints, calculate the fitness value of each individual in the population, and normalize it.

[0096] Step 17): After selecting the best individual in the current population based on the fitness value of each individual, genetic operations are performed on the population. First, a roulette wheel selection operation is used. Each individual is assigned a probability proportional to its fitness value for selection; individuals with higher fitness are more likely to be selected as offspring. This helps retain individuals with high fitness and helps the population evolve towards better individuals. Next, a crossover operation is performed; a portion of the chromosomes from different parent individuals is exchanged according to the crossover probability Pc to generate new offspring individuals. This helps combine the advantages of different individuals and accelerates the population's evolutionary process. Finally, a mutation operation is performed. The chromosomes in the parent individuals are randomly changed according to the mutation probability Pm. For the binary-coded individuals in Example 1, this is achieved by flipping multiple bit codes.

[0097] Step 18): Following the operations in Step 17), a new generation of the population is generated. The best individual from the previous generation replaces the worst individual in the current population, and then genetic iteration is performed continuously to ensure that more excellent individuals are preserved.

[0098] Step 19): Set the maximum number of iterations. When the termination condition is met, decode the current best individual according to the prescribed encoding method and output the optimal target angle estimate under the current snapshot.

[0099] Step 20), repeat steps 6.2) to 6.6) to calculate the average optimal target angle estimate from multiple snapshots;

[0100] To demonstrate that the method described in this invention outperforms traditional DOA estimation algorithms in the context of sea surface, simulation analysis is presented below, using the root mean square error (RMSE) as the performance evaluation criterion. RMSE is defined as follows:

[0101] ;

[0102] The third embodiment: The relationship between the root mean square error of angle measurement and the signal-to-noise ratio (RMS ratio) of the method proposed in this patent and other traditional angle measurement methods under a sea surface background is studied. The resulting curve is shown in the figure below. Figure 4 As shown. The simulation parameters set are as follows:

[0103] Radar array element count: 15, element spacing: 1 meter (half wavelength of the signal). Number of snapshots: 20. Spatial sampling angle range: [-10, 0], number of samples: 100. Target arrival angle: 1.81°. Number of arrivals in both specular and diffuse reflection regions: 3 each, -5.5°, -4.27°, -3.2° and -3.1°, -1.51°, -1.21° respectively. Disturbance coefficient. The population size is 100 individuals, with a mutation probability of 0.05, a crossover probability of 0.9, and a maximum number of iterations of 100. The chromosome length is 8x16. The signal-to-noise ratio range is [0dB:20dB], with a step size of 2dB. 100 independent Monte Carlo experiments were conducted using MATLAB. Simulation results are as follows: Figure 4 As shown.

[0104] according to Figure 4 It is evident that the angle measurement performance of all algorithms improves with increasing signal-to-noise ratio (SNR). The angle measurement performance of the algorithm proposed in this patent gradually plateaus after the SNR exceeds 10 dB. Overall, the algorithm proposed in this patent outperforms other algorithms. This is because the SS-MUSIC, AP-MUSIC, and POMP algorithms are all based on the traditional specular reflection multipath model, which suffers from severe mismatch in a marine environment. Therefore, their angle measurement performance is relatively poor.

Claims

1. A DOA estimation method based on multipath equivalence in the sea surface scattering region, characterized in that, The method includes the following: S1: Set up the radar array to sample the received signal; the radar array is a uniform linear array composed of M array elements with an element spacing of md, where m = 0, 1, 2, ..., M-1, and d is half the wavelength of the incident signal. Considering only the case of a single target incoming wave, the signal data vector received by the radar at time t is: ,in This indicates the direct wave component of the received signal. This represents the received multipath signal components, where P1 and P2 represent the number of incoming multipath signals in the specular reflection region and the number of incoming multipath signals in the diffuse reflection region, respectively. In the formula Represents the transpose of a matrix. , and These represent the direct wave angle steering vector, the specular reflection angle steering vector, and the diffuse reflection angle steering vector, respectively. and Let be the reflection coefficients of the i-th multipath signal from the specular reflection region and the j-th multipath signal from the diffuse reflection region, respectively. Let n(t) be the real envelope of the narrowband signal source incident on the uniform linear array of the radar, and let n(t) represent the additive white Gaussian noise vector received by the array. S2: Based on the spatial coherence of adjacent multipath signals, the multipath signals from the specular reflection region and the multipath signals from the diffuse reflection region are processed equivalently to construct an equivalent multipath signal model of the sea surface. S3: Verify the effectiveness of the equivalent multipath model based on compressed sensing and least squares methods; S4: Based on the sea surface multipath equivalent model, reasonable constraints are imposed on the parameters to be determined; S5: Based on the equivalent multipath model of the sea surface and the conditional constraints obtained in S4, the DOA estimation problem of low elevation angle targets on the sea surface is transformed into a conditional multi-parameter optimization problem. S6: Use a genetic algorithm to solve a multi-parameter optimization problem with constraints. The cost function is converted into a fitness function, and the population is initialized with binary encoding. The parameters to be solved are converted from binary to decimal within the feasible region, and the fitness value of each individual in the population is calculated. Genetic operations are performed. Superior individuals from the original population are introduced into the next generation of the newly generated population to update the population. This process is repeated iteratively until the termination condition is met, and the optimal result is output. The specific details of S2 are as follows: S201: The raw signal received by the radar Normalization is performed to obtain the normalized signal vector; specifically: , in, This represents the maximum value among the vector elements after taking the modulo. S202: Based on the concept of multipath vector synthesis, using and The normalized multipath components of multiple specular reflection regions and multiple diffuse reflection multipath components in S201 are replaced respectively, and an equivalent multipath signal model is constructed: ， in, Normalized noise that includes equivalent fitting error; It is the virtual equivalent reflection coefficient of the direct waveguide vector; and These are the complex equivalent reflection coefficients for the specular reflection region and the diffuse reflection region, respectively.

2. The DOA estimation method based on multipath equivalence in the sea surface scattering region as described in claim 1, characterized in that, The specific details of S3 are as follows: S301: Set T sampling angles in the airspace. Construct a complete dictionary matrix ; S302: Constructing an optimized function based on an overcomplete dictionary: ， in, This is the vector sum of multiple multipath signals in the mirror region or diffuse reflection region of the sea surface. The L0 norm of a vector. This represents a sparse vector to be determined containing only one non-zero element; S303: Subtract each column vector of the overcomplete matrix from the vector sum of the multipath signals. Performing least squares calculations, we obtain the following formula: ， i=1,...,T, where... represent The only non-zero element in; S304: Substitute the results obtained from S303 into S302 to find the minimum value. The sampling angle corresponding to this minimum value is the optimal equivalent multipath angle.

3. The DOA estimation method based on multipath equivalence in the sea surface scattering region as described in claim 2, characterized in that, The specific details of S4 are as follows: S401: Compare and match the normalized expression in S201 with the equivalent multipath signal model in S202 to obtain: ； S402: Using the trigonometric Cauchy inequality, simplifying the expression in S401, we get: ， The amplitude range of each equivalent multipath reflection coefficient can be limited to [0,1], and the phase can be set to... Within the range; S403: Since DOA estimation occurs after target detection, the incoming target signal only exists within half the radar's 3dB beamwidth, thus limiting the direct wave angle to... ,in This represents the radar's 3dB beamwidth. S404: Since the multipath region of specular reflection is located near the specular reflection angle, the source from the mirror can also be set to... ; S405: Since the multipath region of specular reflection is located near the specular reflection angle, and the optimal equivalent specular reflection angle is located within the multipath region, the equivalent specular reflection angle can be set to... ; S407: The diffuse multipath region is larger than the specular multipath region, but multipath echo signals exceeding half the width of the radar's first null beamwidth can be ignored, and the optimal equivalent diffuse reflection angle is also within the diffuse multipath region. Therefore, the equivalent specular reflection angle can be set to... ,in This represents the first null beamwidth of the radar.

4. The DOA estimation method based on multipath equivalence in the sea surface scattering region as described in claim 3, characterized in that, The specific details of S5 are as follows: S501: Combining S2 and S4, the constrained multi-parameter optimization problem is as follows: 。 5. The DOA estimation method based on multipath equivalence in the sea surface scattering region as described in claim 4, characterized in that, The details of S6 are as follows: S601: Based on the genetic algorithm, the fitness function is constructed as follows: ， Each individual in the population corresponds to 8 search parameters, and each individual in the population is represented as follows: ， in Let N be the i-th individual in the population, and N be the number of individuals in the population. These are the direct wave angle, the mirror-equivalent multipath angle, and the diffuse-equivalent multipath angle, respectively. This represents the amplitude and phase of the specular equivalent reflection coefficient. This represents the amplitude and phase of the equivalent reflection coefficient. The equivalent reflection coefficient of the direct wave; S602: Using binary encoding, each parameter in an individual is sequentially concatenated with 16-bit binary code to form a decoded individual; S603: For each individual in the population, perform decimal decoding within the constraints, calculate the fitness value, and normalize it; S604: Based on the fitness value of each individual in the population, select the best individual in the current population; The individuals in the original population are subjected to roulette wheel selection, uniform crossover, and mutation operations in sequence; S605: The best individuals from the previous generation are retained and replaced with the worst individuals in the current population. Then, the genetic iteration is continuously performed so that more excellent individuals are retained. S606: Set the maximum number of iterations. When the termination condition is met, decode the current best individual according to the specified encoding method and output the optimal target angle estimate under the current snapshot. S607: Repeat S602 to S606 to calculate the average optimal target angle estimate from multiple snapshots.

Images