Cylindrical array optimization and sound field reconstruction method based on genetic algorithm

Abstract

The application discloses a kind of based on genetic algorithm's cylindrical surface hydrophone array optimization and sound field reconstruction method, belong to underwater sound field measurement and reconstruction technical field, this method includes: by simulation and equivalent source method, calculate the relative reconstruction error based on real sound pressure and equivalent sound pressure on cylindrical surface;Randomly initialize genetic algorithm population and repair the minimum spacing of constrained hydrophone;In the main loop of genetic algorithm, with the multi-objective fitness function including reconstruction error, space coverage, uniformity index and spacing penalty drives optimization, the population is iteratively updated by selection, crossover, mutation and elite reservation, and then the optimal hydrophone array layout is output;Based on the optimal hydrophone array layout, sound pressure is collected and L1 regularization sparse equivalent source method is used to complete sound field reconstruction.The application deeply couples array layout and reconstruction algorithm, minimizes reconstruction error under fixed number of hydrophones, considers array physical realizability, and significantly improves underwater sound field reconstruction accuracy.

Description

Technical Field

[0001] This invention belongs to the field of underwater sound field measurement and reconstruction technology, and relates to a method for optimizing cylindrical hydrophone arrays and reconstructing sound fields based on genetic algorithms. Background Technology

[0002] Sound field reconstruction aims to invert the sound pressure distribution across an entire space or a specified reconstruction surface using sound pressure data from a limited number of measurement points. It has important applications in noise source identification, acoustic holography, and virtual reality. The Equivalent Source Method (ESM) is a commonly used sound field reconstruction technique. It utilizes a series of virtual equivalent sources located within the sound source region to represent the actual sound source. By measuring the sound pressure on the surface, it deduces the equivalent source intensity and then calculates the sound pressure on the reconstruction surface. Existing ESM-based sound field reconstruction methods typically employ regular arrays, such as uniformly spaced rectangular grids, equiangular cylindrical arrays, or random arrays, to collect sound pressure data.

[0003] Common hydrophone arrays include: Planar arrays: Hydrophones are uniformly distributed on a plane, suitable for near-field acoustic holography, but have limited measurement capabilities for surround sound sources. Spherical arrays: Suitable for three-dimensional spatial sound field analysis, but complex to manufacture and costly. Cylindrical arrays: Balancing azimuth resolution and height coverage, commonly used for pipe sound fields and rotating machinery noise measurement.

[0004] Although existing cylindrical hydrophone arrays and their corresponding ESM reconstruction methods can achieve a certain degree of sound field reconstruction, the following problems still exist:

[0005] Mismatch between array layout and reconstruction mechanism: Traditional uniform arrays do not consider the spatial distribution of equivalent sources and the directional characteristics of sound field propagation, resulting in redundant or insufficient information acquisition and large reconstruction errors.

[0006] Poor adaptability to sparse sound sources: When the number of real sound sources is small and the space is sparse, the uniform array cannot make full use of this sparsity, and the reconstruction results are easily affected by noise amplification.

[0007] Balancing computational efficiency and accuracy is challenging: increasing the number of hydrophones can improve accuracy, but it also increases hardware costs and computational burden. Existing methods lack systematic optimization techniques to maximize reconstruction accuracy with a fixed number of hydrophones.

[0008] Lack of handling of array geometric constraints: In actual hydrophone installations, there are minimum spacing constraints to avoid physical interference and spatial overlap, and traditional optimization methods often ignore or have difficulty handling these constraints.

[0009] In recent years, some studies have attempted to use sparse array optimization methods, such as random sampling based on compressed sensing and sensor layout based on mutual information criteria. However, these methods either require solving complex optimization problems or do not directly use the sound field reconstruction error as the optimization objective.

[0010] Genetic algorithms (GA), as global optimization tools, have been applied in sensor layout optimization, but there is still no systematic solution for the specific problem of underwater acoustic field reconstruction based on sparse measurements of cylindrical arrays. Summary of the Invention

[0011] To address the problem of low sound field reconstruction accuracy caused by unreasonable cylindrical hydrophone array layout in existing technologies, this invention proposes a method for optimizing cylindrical hydrophone arrays and reconstructing sound fields based on a genetic algorithm. By directly incorporating the reconstruction error into the fitness function and introducing array coverage, uniformity, and spacing constraints, the optimal hydrophone distribution is obtained, thereby minimizing the sound pressure error on the reconstruction surface with a fixed number of hydrophones. The specific technical solution is as follows:

[0012] A method for optimizing and reconstructing the sound field of a cylindrical hydrophone array based on a genetic algorithm includes: calculating the relative reconstruction error on the cylindrical surface based on the actual sound pressure and the equivalent sound pressure using simulation and the equivalent source method; randomly generating an initial population containing several individuals based on the genetic algorithm, each individual consisting of coordinate pairs of hydrophones on the cylindrical surface with several repair constraints; performing iterative iterations of the genetic algorithm: for each generation, optimization is driven by a fitness evaluation that introduces the relative reconstruction error, and the population is iteratively updated through tournament selection, crossover and repair, mutation and repair, and elite retention operations; the fitness evaluation function comprehensively considers the hydrophone spatial coverage, spacing uniformity, and spacing penalty; the repair constraints are used to check and repair the angle and height of the hydrophones; selecting the individual with the highest fitness from the final population obtained after iterative updates, and outputting the optimal hydrophone array layout; and reconstructing the sound field based on the optimal hydrophone array layout.

[0013] Furthermore, the calculation of the relative reconstruction error specifically involves: converting the hydrophone's cylindrical coordinates to Cartesian coordinates; randomly generating a real sound source near the simulated equivalent source surface; generating high-resolution grid points on the reconstructed cylindrical surface and calculating the sound pressure generated by the real sound source at these points to obtain the real sound pressure; generating several candidate positions on the holographic cylindrical surface and calculating the ideal noise-free pressure; obtaining the sound pressure at the hydrophone through nearest neighbor interpolation and adding Gaussian white noise to simulate measurement noise; using the sparse equivalent source method to solve for the equivalent source intensity; calculating the equivalent sound pressure on the reconstructed cylindrical surface based on the equivalent source intensity; and using the real sound pressure to calculate the relative reconstruction error ε.

[0014] Furthermore, the equivalent source strength is solved using the L1 regularized sparse equivalent source method.

[0015] Furthermore, the fitness assessment specifically includes:

[0016] Calculate spatial coverage C: Divide the angle-height coordinate plane into grids and count the proportion of grids covered by at least one hydrophone;

[0017] Uniformity U is calculated by weighting the standard deviation of the angle interval, the standard deviation of the height interval, and the corresponding ideal angle interval and height interval.

[0018] Calculate the spacing penalty P: For hydrophone pairs that violate the minimum angular spacing and minimum height spacing constraints, accumulate the penalty;

[0019] Calculate fitness , where δ is a value to prevent division by zero.

[0020] Furthermore, after the crossover and mutation operations, repair constraints are applied to individuals in each generation.

[0021] Furthermore, the repair constraint specifically involves: traversing all hydrophones and checking whether their angle difference and height difference are less than the preset minimum angle interval and minimum height interval, respectively. If so, they are translated along the circumference to the minimum angle interval and minimum height interval; otherwise, the original angle and height are retained.

[0022] Furthermore, the crossover operation adopts uniform crossover: an M-dimensional binary mask is generated. When the mask is 1, offspring 1 takes the corresponding gene of parent 1, and offspring 2 takes the corresponding gene of parent 2; when the mask is 0, they are swapped.

[0023] Furthermore, the mutation operation involves mutating a random proportion of hydrophones in the offspring individuals with a mutation probability, applying random perturbations to their azimuth angle coordinates, and applying random perturbations to their height coordinates that are proportional to the height range of the cylindrical surface.

[0024] Furthermore, the sound field reconstruction based on the optimal hydrophone array layout specifically involves: installing hydrophones in an optimal array on a holographic cylindrical surface to collect sound pressure signals, using a sparse equivalent source reconstruction method to solve for the equivalent source intensity, calculating the sound pressure distribution on the reconstructed cylindrical surface, and completing the sound field reconstruction.

[0025] The beneficial effects of this invention include:

[0026] 1. Genetic optimization based on reconstruction error: The relative error of reconstruction by the equivalent source method is directly used as the core part of the fitness function of the genetic algorithm, so that the array layout and reconstruction algorithm are deeply coupled to achieve end-to-end performance optimization.

[0027] 2. The design of a multi-objective fitness function takes into account reconstruction error, spatial coverage, angle-height uniformity and minimum spacing constraint, and balances reconstruction accuracy with array physical feasibility.

[0028] 3. A repair operator is proposed to explicitly handle the geometric constraints of the cylindrical surface. After crossing and mutation, the minimum angular interval and minimum height interval are forced to be satisfied, ensuring that the arrays corresponding to all individuals can be installed in the actual system.

[0029] 4. Introduce angle and height uniformity indices. The uniformity index is a regularization term in the fitness function of the genetic algorithm. It is used to guide the algorithm to generate a hydrophone array that is both accurate and spatially balanced, avoiding the sacrifice of sampling uniformity for the reduction of small errors, thereby improving the overall performance and practical usability of sound field reconstruction.

[0030] 5. Sparse equivalent source reconstruction and array optimization synergy: The L1 regularized sparse reconstruction method is adopted, which forms a synergistic effect with genetic optimization, further improving the reconstruction capability of sparse sound sources.

[0031] 6. Strong generalization ability: It is applicable to different frequencies, different cylindrical surface sizes, different numbers of hydrophones and different sound source distributions. By adjusting the genetic algorithm parameters, it can quickly adapt to new scenarios. Attached Figure Description

[0032] Figure 1 The overall flowchart of the method of this invention includes steps such as data input, genetic algorithm iteration, optimal array output, and sound field reconstruction.

[0033] Figure 2 : Geometric schematic diagram of holographic cylindrical surface, reconstructed cylindrical surface and equivalent source surface, including the location of equivalent source and real sound source; In the figure: 101-holographic cylindrical surface, which is also the hydrophone mounting surface; 102-reconstructed cylindrical surface; 103-equivalent source surface; 104-real sound source; 105-equivalent source.

[0034] Figure 3 The chromosome encoding method of individuals in the population in the genetic algorithm is shown. Each individual consists of a chromosome structure composed of M (θ, z) pairs. Each individual corresponds to a set of cylindrical hydrophone array layouts and is encoded as a fixed-length real number vector.

[0035] Figure 4 : A diagram illustrating the crossover operation, showing a uniform crossover.

[0036] Figure 5 : A schematic diagram of mutation operations, illustrating random perturbation mutations.

[0037] Figure 6 The flowchart for repairing constraint operators demonstrates how to adjust the position of hydrophones that violate the minimum spacing constraint.

[0038] Figure 7 The layout of the hydrophone array optimized by this invention on a cylindrical surface is shown in the figure, and compared with random arrays and regular arrays.

[0039] Figure 8 Comparison of sound pressure amplitude distribution on the reconstructed cylindrical surface, where: (a) theoretical sound pressure on the reconstructed surface; (b) reconstructed sound pressure of the hydrophone array optimized in this invention on the reconstructed surface.

[0040] Figure 9 The convergence curve of the genetic algorithm shows how fitness and reconstruction error change with the number of iterations.

[0041] Figure 10 Histogram comparison of different array layouts in terms of reconstruction error, minimum spacing, and uniformity performance. Detailed Implementation

[0042] To make the objectives, technical solutions, and technical effects of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0043] like Figure 1 As shown, this embodiment provides a method for optimizing and reconstructing the sound field of a cylindrical hydrophone array based on a genetic algorithm, including the following steps:

[0044] Step 1: Initialize system parameters.

[0045] The following parameters are either input by the user or preset by the system:

[0046] Acoustic parameters: frequency f, sound velocity c0, water density ρ0.

[0047] Geometric parameters: radius of the cylindrical holographic surface Reconstructing the radius of the cylinder surface Equivalent source surface height Equivalent source surface size Equivalent source grid number (Quantity per side), Cylinder height range ~ .

[0048] Number of hydrophones: M.

[0049] Position of the i-th hydrophone on the cylindrical holographic surface Reconstruct the position of the i-th hydrophone on the cylindrical surface. The position of the j-th equivalent source on the equivalent source surface .

[0050] Genetic algorithm parameters: population size Maximum number of iterations K, crossover probability Probability of mutation The number of elites retained is E.

[0051] Constraint parameter: Azimuth Minimum angular interval Minimum height interval .

[0052] Reconstruction parameters: sparse regularization parameter λ, noise level estimate ϵ.

[0053] Step 2: Generate a real sound source 104 and reference data for simulation evaluation.

[0054] like Figure 2 As shown, random generation occurs near the equivalent source surface 103. There are 104 real sound sources, each with a complex intensity qs.

[0055] Generate high-resolution mesh points on the reconstructed cylindrical surface 102, for example , Calculate the sound pressure generated by the real sound source 104 at these points. This serves as the benchmark for calculating reconstruction errors.

[0056] A large number of candidate locations, such as 2000 random points, are generated on the holographic cylindrical surface 101. The ideal noise-free sound pressure is calculated for rapid sampling in subsequent fitness evaluation.

[0057] Step 3: Initialize the genetic algorithm parameter population.

[0058] like Figure 3 As shown, Np individuals are generated, each individual is an M×2 matrix, and the matrix is ​​stored... Yes, the radius is fixed. For each individual's M hydrophones, randomly generate... .

[0059] Call the repair constraint operator to adjust the position of the hydrophone until the minimum angle and height interval constraints are met.

[0060] The current population is denoted as Algebraic counter .

[0061] Step 4: Main loop of the genetic algorithm.

[0062] Continue to refer to Figure 1 For each generation Perform the following sub-steps:

[0063] Step 4.1, Fitness Assessment: For the population... For each individual x corresponding to a set of hydrophone coordinates, execute:

[0064] Step 4.1.1: Convert the hydrophone cylindrical coordinates to Cartesian coordinates.

[0065] Step 4.1.2: Based on the ideal sound pressure at the candidate location, obtain the sound pressure at the hydrophone through nearest neighbor interpolation. , ( ), where the superscript H represents the conjugate transpose operator.

[0066] Step 4.1.3, To Add Gaussian white noise to simulate the measurement of signal-to-noise ratio (SNR).

[0067] Step 4.1.4: Reconstruct using the sparse equivalent source method to obtain the equivalent source intensity q.

[0068] Step 4.1.5: Calculate the sound pressure on the reconstructed cylindrical surface 102. , This represents the sound field transfer matrix.

[0069] Step 4.1.6: Calculate the relative reconstruction error: .

[0070] Step 4.1.7, Calculate spatial coverage C: The plane is divided into Grid, the percentage of grids covered by at least one hydrophone.

[0071] Step 4.1.8: Calculate uniformity ,in , These are the standard deviations of the angle interval and the height interval, respectively, and the ideal height interval. Ideal angle interval .

[0072] Step 4.1.9, Calculate the spacing penalty P: For all hydrophone pairs, calculate the angle difference. and height difference ,like Then the penalties will be accumulated. The same applies to height; finally, divide by the total logarithm to get the average.

[0073] Step 4.1.10: Calculate fitness ,in Prevent division by zero.

[0074] Step 4.2: Record the best individual: Find the individual with the highest fitness and save its reconstruction error and coordinates.

[0075] Step 4.3, Selection: A tournament selection method will be used, with random selection each time. For each individual, the one with the highest fitness is selected to enter the cross-matching pool, and this process is repeated. Next, generate the matching pool index.

[0076] Step 4.4, Crossover: Each hydrophone of the two paired individuals is exchanged with the same crossover probability, thus forming two new individuals. (See reference...) Figure 4 As shown. For two adjacent parent generations in the matching pool, with probability... Perform uniform crossover: Generate an M-dimensional binary mask. When the mask is 1, child generation 1 takes the value of parent generation 1. Child generation 2 takes the parent generation 2. When the mask is 0, swap. After crossing, call the constraint repair operator. If not crossing, the child directly copies the parent.

[0077] 4.5. Variation: such as Figure 5 As shown, new individuals are formed by replacing the gene values ​​at certain gene loci in the coding sequence of an individual's chromosome with other alleles at those loci. For each offspring individual, a probability is applied... Mutation: Randomly select 10% of the hydrophones and add θ + 0.5 to their respective values. Add z+0.1 to z. Then, the constraint repair operator is invoked.

[0078] 4.6 Elite Preservation: The E individuals with the highest fitness in the current population are directly replaced with the individuals with the lowest fitness in the offspring population. Individual.

[0079] 4.7 Update the population: Update the offspring population Return to step 4.1 until the maximum algebra is reached.

[0080] like Figure 6 As shown, the constraint repair operator in the above steps performs the following: traversing all hydrophone pairs, checking whether their angle difference and height difference satisfy the minimum angle interval and minimum height interval constraints, respectively; for hydrophones that violate the constraints, increasing the offset along the violation direction until the constraints are satisfied. Specifically, the input hydrophone coordinate set... The constraint parameters in the angular and height directions are respectively , Execute the following respectively:

[0081] Angle and direction constraint check: Traverse all hydrophone pairs and calculate =| |, when Translation along the circumference to Conversely, retain the original angle;

[0082] Height constraint check: Traverse all hydrophone pairs and calculate =| |, when Translation along the circumference to Conversely, retain the original height;

[0083] After execution, the individual with the corrected constraints is output, satisfying the constraint parameters. , .

[0084] Step 5: Output the optimal array.

[0085] refer to Figure 7 As shown, after the genetic algorithm completes, the individual with the highest fitness is selected from the final population. This is the optimal hydrophone array layout.

[0086] Step 6: Sound field reconstruction based on the optimal array.

[0087] like Figure 8 As shown, in the actual measurement, the hydrophones were mounted in an optimal array on the holographic cylindrical surface 101 to collect sound pressure signals. .

[0088] Sparse equivalent source reconstruction is performed to obtain the equivalent source intensity, and the sound pressure on the reconstructed cylindrical surface 102 is calculated.

[0089] Output the reconstructed sound pressure distribution to complete the sound field reconstruction.

[0090] Step 7: Performance Comparison.

[0091] To verify the optimization effect, a random array and a regular grid array with equal angle and height intervals can be generated simultaneously. The reconstruction error can be calculated according to steps 4.1.1 to 4.1.6 and compared with the optimal array.

[0092] Corresponding to the aforementioned embodiments of the cylindrical hydrophone array optimization and sound field reconstruction method based on genetic algorithms, this invention also provides an embodiment of a cylindrical hydrophone array optimization and sound field reconstruction system based on genetic algorithms. The system includes:

[0093] The sound field measurement geometry configuration unit is initialized with the following parameters:

[0094] Holographic cylindrical surface 101: radius is The cylindrical surface has a height range of [ , ], used to mount hydrophone arrays. Hydrophones are represented in cylindrical coordinates by ( )express.

[0095] Reconstructing cylindrical surface 102: radius is The cylindrical surface is the target surface from which the sound pressure distribution needs to be reconstructed. Typically... < .

[0096] Equivalent source surface 103: located at height A plane on which are evenly distributed One equivalent source 105 is used to represent an equivalent actual sound source.

[0097] The genetic algorithm optimization unit includes the following sub-modules:

[0098] Initialization module: Generates the initial population, each individual containing M hydrophones on a cylindrical surface ( Coordinates. During initialization, positions are randomly generated and the repair operator module is called to ensure that the minimum angle interval is met. and minimum height interval (m).

[0099] Fitness assessment module: Perform the following calculations for each individual:

[0100] By sampling sound pressure based on the coordinates of the hydrophone, an ideal measurement can be simulated and noise can be added.

[0101] Call the sparse equivalent source reconstruction module to solve for the equivalent source strength.

[0102] Calculate the sound pressure on the reconstructed cylindrical surface and its relative error ε with the actual sound pressure.

[0103] Calculate spatial coverage C, uniformity U, and spacing penalty P.

[0104] Output fitness .

[0105] Selection module: Tournament selection is adopted, with tournament size k=3, and individuals with higher fitness are selected from the current population to enter the mating pool.

[0106] Crossover module: Probably selects parent individuals... Perform uniform crossover: randomly select from parent 1 or parent 2 for each hydrophone location. After the crossover, the constraint repair operator module is called.

[0107] Mutation module: For individuals after crossover, use probability... Mutation is performed: Randomly select 10% of the hydrophones and add random perturbations to their θ and z values. The perturbation amplitude for θ is 0.5 rad, and the perturbation amplitude for z is 0.1 times the height range. After mutation, the repair constraint operator module is called.

[0108] Elite Preservation Module: The E individuals with the highest fitness in the current population are directly copied to the next generation to ensure that the optimal solution is not lost.

[0109] Repair constraint operator module: Considering the periodicity of the circle, check the angle difference and height difference of all hydrophone pairs in the individual. If they are below the threshold, increase the spacing by moving the position and clamp to the effective range.

[0110] Optimal Array Output Module: After the genetic algorithm iteration is completed, the hydrophone coordinates corresponding to the individual with the highest fitness are output as the optimal array.

[0111] A sparse equivalent source reconstruction module is implemented to solve for the equivalent source intensity based on L1 regularization.

[0112] Input: Hydrophone location from the optimal array or the individual to be evaluated, and measured sound pressure level. Equivalent source surface coordinates.

[0113] Construct the transfer matrix between the holographic surface and the equivalent source surface ,element , Among them, subscript .

[0114] Construct the transfer matrix between the reconstructed surface and the equivalent source surface. ,element , .

[0115] In the above formula, i represents the imaginary number sign, k represents the wave number, and k = 2πf / c0.

[0116] The Iterative Soft Thresholding Algorithm (ISTA) is used to solve this problem.

[0117] .

[0118] Output equivalent source strength .

[0119] The sound pressure reconstruction and retrieval matching unit utilizes the reconstructed equivalent source intensity. Calculate the sound pressure at each point on the reconstructed cylindrical surface 102: .

[0120] The connections between modules are as follows: The output of the initialization module is connected to the fitness evaluation module and the constraint repair operator module. The output of the fitness evaluation module is fed to the selection module. The selection module is connected to the crossover module, and the crossover module is connected to the mutation module. The outputs of the mutation module and the elite retention module together form the next generation population, which returns to the fitness evaluation module to form an iterative loop. After the iteration terminates, the optimal array output module outputs the optimal hydrophone coordinates to the sparse equivalent source reconstruction module, which is connected to the sound pressure reconstruction and retrieval matching unit.

[0121] The technical effects of this invention are illustrated below through simulation experiments and analysis:

[0122] Experimental parameters: frequency 500Hz, sound velocity 1500m / s, cylindrical holographic surface radius 0.5m, reconstructed surface radius 0.4m, height range ±0.3m, number of hydrophones 60, equivalent source surface 11×11 grid, 2 real sound sources, signal-to-noise ratio 30dB.

[0123] like Figure 10 As shown, the reconstruction errors of different arrays are: random array, relative reconstruction error is 5.6%; regular grid array, relative reconstruction error is 30.0%; and the optimal array output by this invention has a relative reconstruction error of 3.4%. This demonstrates that this invention can effectively tap the information potential of a limited number of hydrophones.

[0124] This invention optimizes a cylindrical hydrophone array using a genetic algorithm, fully considering the sound field propagation mechanism and array geometric constraints. It significantly improves the accuracy of underwater sound field reconstruction with a fixed hardware cost, while ensuring the physical feasibility of the array and the robustness of the algorithm, thus possessing high practical value. This is explained in detail below:

[0125] like Figure 9 As shown, the convergence of the genetic algorithm of this invention is as follows: the fitness curve and reconstruction error curve converge steadily with the increase of generations, typically reaching more than 90% of the optimal solution within 30 generations, and fully converging after 50 generations. This demonstrates that the genetic algorithm can efficiently search the array layout space.

[0126] Synergistic effect of sparse reconstruction and array optimization: Under L1 regularized sparse equivalent source reconstruction, the reconstruction error of the optimal array of the present invention is reduced by 41% compared with ordinary least squares reconstruction, indicating that the optimized array and the sparse method have a synergistic effect and can better identify the dominant equivalent source.

[0127] Frequency robustness: In tests within the range of 200 Hz to 2000 Hz, the reconstruction error of the optimal array of this invention is consistently lower than that of the comparison array, especially in the low-to-mid frequency range (200-800 Hz), where the improvement is most significant (error reduction of 25%-35%). In the high-frequency range (>1500 Hz), the improvement is slightly reduced but still maintains its advantage.

[0128] Generalization ability: When the method of the present invention was applied to different numbers of real sound sources (1 to 5), different equivalent source grid densities (9×9, 11×11, 15×15), and different numbers of hydrophones (30, 60, 90), a consistent improvement in reconstruction accuracy was observed, indicating that the optimization framework of the present invention has good generalization ability.

[0129] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any way. Although the implementation process of the present invention has been described in detail above, those skilled in the art can still modify the technical solutions described in the foregoing examples or make equivalent substitutions for some of the technical features. All modifications and equivalent substitutions made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for optimizing and reconstructing the sound field of a cylindrical hydrophone array based on a genetic algorithm, characterized in that, include: The relative reconstruction error ε based on the real sound pressure and the equivalent source method on the cylindrical surface is calculated using simulation and the equivalent sound pressure. , Equivalent sound pressure level, This is the actual sound pressure level; Based on the genetic algorithm, an initial population containing several individuals is randomly generated. Each individual consists of several pairs of coordinates of hydrophones with repair constraints on the cylindrical surface. The genetic algorithm is executed iteratively: for each generation, optimization is driven by a fitness evaluation function that introduces the relative reconstruction error, and the population is iteratively updated through tournament selection, crossover and repair, mutation and repair, and elite retention operations; The fitness evaluation function comprehensively considers hydrophone spatial coverage, spacing uniformity, and spacing penalty; Calculate spatial coverage Divide the angle-height coordinate plane into a grid and count the percentage of grids covered by at least one hydrophone; Calculate uniformity Based on the weighted synthesis of the standard deviation of the angle interval, the standard deviation of the height interval, and the corresponding ideal angle and height intervals, the calculation expression is as follows: ,in , These are the standard deviations of the angle interval and the height interval, respectively, and the ideal height interval. Ideal angle interval ; It is the number of hydrophones. and These are the maximum and minimum values ​​within the height range of the cylindrical surface, respectively. Calculate spacing penalty For hydrophone pairs that violate the minimum angle and minimum height spacing constraints, an accumulated penalty is applied. Specifically, for all hydrophone pairs, the angle difference is calculated. and height difference ,like Then the penalties will be accumulated. , This is the minimum angular interval, and the same applies to height; finally, divide by the total logarithm to get the average value. Calculate fitness , where δ is the value to prevent division by zero; The repair constraint is to check and repair the angle and height of the hydrophone. Specifically, it involves: traversing all hydrophone pairs and checking whether their angle difference and height difference are less than the preset minimum angle interval and minimum height interval, respectively. If so, it is translated along the circumference to the minimum angle interval and minimum height interval; otherwise, the original angle and height are retained. Select the individual with the highest fitness from the final population obtained after iterative updates, and output the optimal hydrophone array layout; Sound field reconstruction is performed based on the optimal hydrophone array layout.

2. The method for optimizing and reconstructing the sound field of a cylindrical hydrophone array as described in claim 1, characterized in that, The calculation of the relative reconstruction error is as follows: Convert the hydrophone cylindrical coordinates to Cartesian coordinates; A real sound source is randomly generated near the equivalent source surface in the simulation. High-resolution grid points are generated on the reconstructed cylindrical surface, and the sound pressure generated by the real sound source is calculated at these points to obtain the real sound pressure. Several candidate positions are generated on the holographic cylindrical surface and the ideal noise-free pressure is calculated. The sound pressure at the hydrophone is obtained by nearest neighbor interpolation and Gaussian white noise is added to simulate the measurement noise. The sparse equivalent source method is used to solve the equivalent source intensity. The equivalent sound pressure on the reconstructed cylindrical surface is calculated based on the equivalent source intensity, and the relative reconstruction error ε is calculated using the actual sound pressure.

3. The method for optimizing and reconstructing the sound field of a cylindrical hydrophone array as described in claim 2, characterized in that, The equivalent source strength is solved by using the L1 regularized sparse equivalent source method.

4. The method for optimizing and reconstructing the sound field of a cylindrical hydrophone array as described in claim 1, characterized in that, After the crossover and mutation operations, a repair constraint operation is performed on individuals in each generation.

5. The method for optimizing and reconstructing the sound field of a cylindrical hydrophone array as described in claim 1, characterized in that, The crossover operation adopts uniform crossover: an M-dimensional binary mask is generated. When the mask is 1, offspring 1 takes the corresponding gene of parent 1, and offspring 2 takes the corresponding gene of parent 2; when the mask is 0, they are swapped.

6. The method for optimizing and reconstructing the sound field of a cylindrical hydrophone array as described in claim 1, characterized in that, The mutation operation involves mutating a random proportion of hydrophones in offspring individuals using a mutation probability, applying random perturbations to their azimuth angle coordinates, and applying random perturbations to their height coordinates that are proportional to the height range of the cylindrical surface.

7. The method for optimizing and reconstructing the sound field of a cylindrical hydrophone array as described in claim 1, characterized in that, The sound field reconstruction based on the optimal hydrophone array layout is specifically as follows: hydrophones are installed in the optimal array on the holographic cylindrical surface to collect sound pressure signals, the equivalent source intensity is solved by the sparse equivalent source reconstruction method, the sound pressure distribution on the reconstructed cylindrical surface is calculated, and the sound field reconstruction is completed.

Images