Universal sonar signal simulation system and its application

Abstract

A universal sonar signal simulation method is provided, comprising: obtaining a form X of a target to be simulated, obtaining a target sound source feature Y corresponding to the form of the target using a logistic regression prediction model; obtaining the positions of the target and a sonar used to acquire the target in earth coordinates; obtaining, based on the positions of the earth coordinates, a sea area in which the target and the sonar are located, as well as sea area environmental data of the sea area; generating a sea area environmental sound field model of the sea area based on the sea area environmental data; obtaining configuration information of the sonar, and calculating data representing a simulated signal of the target sound received by the sonar based on the sea area sound field model and the target sound source feature.

Description

Technical Field

[0001] The present application belongs to the field of underwater acoustic detection simulation technology, and in particular to a general sonar signal simulation system and its application. Background Art

[0002] The demand for upgrades and functional iterations of underwater acoustic detection equipment is increasing rapidly. Sonar systems, as a crucial component of underwater acoustic detection equipment, undergo a lengthy development process, and performance verification requires extensive testing, including lake and sea trials. Sonar signal simulation systems provide the primary data source for debugging and performance testing of sonar systems. They can significantly shorten sonar development cycles and reduce the number of lake and sea trials. Therefore, sonar signal simulation technology is of great significance.

[0003] Traditional sonar signal simulation systems suffer from complex structures, poor reliability, and a narrow adjustment range. Furthermore, each new sonar device requires a completely new design and development. Traditional sonar signal simulation systems also barely reflect changes in the ocean environment, a highly variable factor that significantly impacts sonar device performance. Therefore, a universal sonar signal simulation system is desired to address these challenges.

[0004] Existing ocean data assimilation technology combines ocean data obtained from different spatial, temporal, and observational methods with mathematical models to provide environmental data (temperature, salinity, depth, wind speed, etc.) for a specific ocean area. The Modular Ocean Data Assimilation System (MODAS) is the primary tool used by the US Navy to determine the three-dimensional temperature and salinity fields of the global ocean. MODAS assimilates sea surface temperature and height data from satellite remote sensing to produce a dynamic climate state that more closely reflects actual ocean conditions (https: / / journals.ametsoc.org / view / journals / atot / 19 / 2 / 1520-0426_2002_019_0240_tmodas_2_0_co_2.xml). There are also several domestic providers of ocean data assimilation based on satellite remote sensing data. Ocean data assimilation makes it possible to simulate the environmental characteristics of specific ocean areas and, in turn, support the simulation of the impact of specific ocean environments on sonar signals. Summary of the Invention

[0005] In view of the fact that the existing technology does not integrate marine environmental information, the purpose of the present invention is to propose an acoustic field simulation solution based on ocean data assimilation data from satellite remote sensing data, so as to solve the problem of introducing the influence effects of environmental factors in sonar signal simulation.

[0006] In view of the fact that the existing technology has not achieved the fusion of target sound source signals with ocean propagation effects, the purpose of the present invention is to propose a signal simulation solution based on adaptive fitting of multiple sound field models to solve the problem of sonar signal simulation under the sound field propagation effect where the target signal characteristics are affected by environmental factors.

[0007] In view of the fact that the existing technology fails to realize the comprehensive simulation of target characteristics, the purpose of the present invention is to propose a solution based on feature analysis, extraction, fitting, approximation and superposition simulation to solve the problem of reconstructing the characteristic components in the target signal.

[0008] In view of the fact that the existing technology has not achieved the integrated configuration of universal sonar platform parameters, the purpose of the present invention is to propose a solution based on a universal integrated parameter interactive interface to solve the problem of integrated configuration of multiple related parameters that affect signal generation, such as base array formation, array element spacing, operating frequency band, signal processing algorithm, etc. for different sonar platforms.

[0009] In view of the high cost of existing technologies, the purpose of this invention is to propose a signal simulation application solution based on a GPU heterogeneous parallel computing hardware platform to solve the conflict between the computing power and hardware cost of traditional signal processors.

[0010] To address the above-mentioned issues, the present invention aims to propose a universal sonar signal simulation system. Based on a universal modular design, this system integrates information from multiple influencing factors of the ocean environment, simulates characteristic-level signals for specific targets, and integrates the combined effects of ambient sound field propagation. It also provides a universal sonar integrated configuration platform and utilizes parallel computing hardware to simulate sonar array-domain and beam-domain signals.

[0011] According to an embodiment of the present application, a general sonar signal simulation method according to the first aspect of the present application is provided, comprising: obtaining the form X of the target to be simulated, using a logistic regression prediction model to obtain a target sound source feature Y corresponding to the form of the target, wherein W TThe result of X+b is used as the prediction of the target sound source feature Y, and W and b are weight vectors of the logistic regression prediction model; the position of the target and the sonar that collects the target in the earth coordinates is obtained; the sea area where the target and the sonar are located and the sea area environment data of the sea area are obtained according to the position of the earth coordinates; a sea area environment sound field model of the sea area is generated according to the sea area environment data, wherein the sea area is divided into one or more sub-sea areas according to the spatial distribution of the sea area environment data, and a corresponding sound field model is matched for each sub-sea area, and the sound field models of each sub-sea area of ​​the sea area together constitute the sound field model of the sea area; the configuration information of the sonar is obtained, and the propagation effect of the target sound in the sea area is calculated according to the sound field model of the sea area and the target sound source feature, and the data of the analog signal representing the target sound received by the sonar is calculated in combination with the configuration information of the sonar.

[0012] According to a second aspect of the present application, a computing device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the general sonar signal simulation method provided in accordance with the first aspect of the present application is implemented.

[0013] The positive and progressive effects of the present invention are: (1) introducing the ocean environment factors that have the most complex and changeable impact on sonar signals and cannot be described by specific physical formulas into the simulation system; (2) making full use of multiple sound field models to perform adaptive signal matching simulation, so as to be as close to the actual propagation effect of the signal as possible; (3) using a fitting approximation method for target feature simulation, and using logistic regression fitting methods to achieve feature signal approximation processing from information such as actual ships and their noise signals, so as to fit the target features to the greatest extent; (4) comprehensively utilizing parallel computing platforms to realize the possibility of parallel calculation of complex parameters of multiple array elements; (5) designing a multi-functional interface interaction and a three-dimensional visualization display effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments recorded in this application. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0015] Figure 1 The overall block diagram of the universal sonar signal simulation system according to an embodiment of the present application is shown.

[0016] Figure 2 A schematic diagram of calculating a sea area environment sound field model according to an embodiment of the present application is shown.

[0017] Figure 3A schematic diagram of the calculation process of fitting target features by the target feature extraction module according to an embodiment of the present application is shown.

[0018] Figure 4 A schematic diagram showing parallel computing of simulated sonar signals by a computing platform according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0019] The following is a clear and complete description of the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0020] Figure 1 The overall block diagram of the universal sonar signal simulation system according to an embodiment of the present application is shown.

[0021] The general sonar simulation system includes a target feature extraction module, a sea area sound field simulation module, a user interaction interface and a signal processing platform.

[0022] The universal sonar signal simulation system simulates the sonar signal of a designated target (for example, a specific ship) collected by a sonar at a designated location under the influence of a designated sea area environment.

[0023] The target feature extraction module is used to simulate the sound source characteristics of the target. To improve versatility and enable the universal sonar signal simulation system of the present embodiment to simulate the sound source characteristics of different targets, the target extraction module generates data such as frequency, bandwidth, and soundprint spectrum based on the target's characteristic parameters (such as shape and speed, provided by the user through a user interface, for example), and provides it to the signal processing platform for simulating sonar signals.

[0024] The sea area sound field simulation module is used to establish an environmental sound field model of the sea area specified by the user. The user provides the location of the specified sea area through the user interaction interface. Optionally, the user also provides the time when the simulation occurs. The environment of the sea area changes over time, thereby affecting the characteristics of the sea area sound field. The sea area sound field simulation system uses the ocean data assimilation system to construct a sea area environmental sound field model based on the historical ocean data, quasi-real-time ocean data and / or forecast ocean data of the specified sea area. The sea area environmental sound field model is also provided to the signal processing platform for simulating sonar signals. Through the user interaction interface, the sea area sound field simulation module also visually displays the three-dimensional sound field of the specified sea area to the user.

[0025] The signal processing platform superimposes the simulated signal under the dual effects of the target's sound source characteristics and the sea area environmental sound field model to simulate the target sonar signal collected by the sonar (including, for example, array domain data and beam domain data).

[0026] The user interface is used for user operations, obtaining parameters of the target sound source provided by the user, obtaining environmental parameters such as the sea area, sea area environment / time, and the sonar configuration information to be simulated provided by the user. The obtained parameters and information are provided to each module, and the signal processing is performed to simulate the sonar signal, and the simulation results are displayed to the user.

[0027] In order to simulate the sonar signal of the target, the user provides the sea area where the target is located, and obtains the terrain data of the selected sea area. For example, the user provides the latitude and longitude information of the sea area, or selects the sea area through the digital earth system. A variety of maps, digital earth, high-definition earth systems, and virtual reality systems provide the terrain data of the earth. The terrain data includes information such as depth, bottom, and topography. The historical ocean data, quasi-real-time ocean data, and / or forecast ocean data provided by the ocean data assimilation system provide information such as temperature, salinity, ocean currents, wind speed, eddies, and ocean fronts of the sea area. According to an embodiment of the present application, data including information such as temperature, salinity, depth, bottom, topography, ocean currents, wind speed, eddies, and ocean fronts are collectively referred to as sea area environmental data. According to the sea area specified by the user, the sea area environmental data of the specified sea area is obtained. And a sea area environmental sound field model is generated based on the sea area environmental data.

[0028] Figure 2 A schematic diagram of calculating a sea area environment sound field model according to an embodiment of the present application is shown.

[0029] Reference Figure 2 The present invention is based on the selection of a sea area, the introduction of the corresponding specific sea area environmental data, and the statistical analysis of the marine environmental data. It mainly conducts a comprehensive evaluation of the sound velocity profile, sound velocity horizontal gradient distribution, thermocline water layer distribution, seabed topography undulation and other information in the selected area, and comprehensively selects the sound field model based on the relative position of the selected target, distance change and the distribution of the main frequency bands of the sound source characteristic information.

[0030] There are various acoustic field models in the prior art, such as the Bellhop model, the Krakan model, and the RAM model. Each of these models has its own advantages and limitations. To expand the applicability of universal sonar signal simulation systems, according to embodiments of the present application, statistical analysis is performed on environmental data from specific selected sea areas, and an adaptive algorithm is employed to implement regional model matching calculations to improve the accuracy of the generated sea area environmental acoustic field model.

[0031] For a specific sea area, statistics are collected on its sound velocity distribution, seabed topography, depth distribution, and other information. The results are then weighted and matched to a suitable acoustic field model. The selected sea area may be adapted to a single model, or it may be divided into multiple sub-areas, each with a different model. The ultimate goal is to calculate the sound pressure field distribution for the entire selected sea area through the acoustic field model.

[0032] According to the selected specific sea area, information such as the horizontal distribution of sound speed, vertical distribution of sound speed, and change rate of seabed elevation data are extracted from its sea area environmental data.

[0033] Construct a sea area grid based on the spatial location of the sea area. A sea area grid is an area enclosed by 0.1° longitude and 0.1° latitude. Therefore, the selected sea area is composed of multiple sea area grids.

[0034] The horizontal sound velocity distribution, vertical sound velocity distribution, and seafloor elevation data change rate for each sea area grid are represented by the same assimilated value. For example, the mean sound velocity within a sea area grid in the sea area environmental data can be used as the assimilated value for the horizontal sound velocity distribution for that sea area grid. Those skilled in the art will appreciate other methods for calculating assimilated values.

[0035] The assimilation values ​​of various information of adjacent sea area grids (for example, the assimilation value of the horizontal distribution of sound speed, the assimilation value of the vertical distribution of sound speed, the assimilation value of the change rate of seabed elevation data, etc.) are compared, and the areas where the assimilation value changes exceed the set threshold are segmented. Finally, the selected sea area is divided into one or more sub-sea areas. Within each sub-sea area, the change in the assimilation value of various information of the adjacent sea area grids of the sub-sea area is less than the set threshold. The change in the assimilation value of the adjacent sea area grids of different sub-sea areas is not less than or greater than the set threshold. Thus, based on the changes in the sea area environmental data, the selected sea area is divided into one or more sub-sea areas.

[0036] The ocean environment data for each sub-sea area (e.g., sound velocity, temperature, topography, salinity, ocean currents, and their distribution) is normalized. Based on the distribution of the ocean environment data for each sub-sea area, a corresponding acoustic field model is selected to construct the acoustic field of the sub-sea area. The propagation effect of sound within the sub-sea area is then determined based on the acoustic field. The propagation effects of sound across all sub-sea areas as it propagates from the target to the sonar location within the selected sea area are accumulated to obtain the sonar signal to be simulated. The acoustic field models for each sub-sea area are accumulated to obtain the overall acoustic field model for the selected sea area.

[0037] According to an embodiment of the present application, for example, a Gaussian ray model is used for a sub-sea area with little change in sound velocity level and relatively flat seabed topography, and a parabolic equation model is used for a sub-sea area with large change in sound velocity level and relatively complex seabed topography. Alternatively, a hierarchical analysis method is used to determine the sound field model corresponding to the sub-sea area:

[0038] (1) Determine the acoustic field model evaluation system and establish a hierarchical correspondence between sub-sea area environmental data and acoustic field models;

[0039] (2) construct a judgment matrix to determine the impact scale of different physical quantities (different attributes of marine environmental data);

[0040] (3) Define the weights corresponding to different physical quantities;

[0041] (4) Calculate the scores for each sub-sea area and each sound field model based on the judgment matrix and weights, and determine the model with the highest score as the sound field model to be applied in a certain sub-area.

[0042] Optionally, a three-dimensional engine is used to visualize the sound field model. And also generates a sound pressure field data file of the sound field model of the selected sea area. Signal processing platform (also see Figure 1 ) According to the sound pressure field data file and the target sound source parameter file, the propagation effect of the target sound in the sound pressure field is calculated to obtain the simulation data of the target sound signal received by the sonar.

[0043] Still taking an example, based on the virtual reality engine UE4 / UE5 (https: / / www.unrealengine.co m / zh-CN), the ocean assimilation database https: / / www.ncei.noaa.gov / products / world-ocean-atlas) and the ocean forecast database (https: / / www.ecmwf.int / en / forecasts) are imported to form sea area environmental data based on three environmental data types: historical ocean environmental data, quasi-real-time ocean environmental data, and forecast ocean data.

[0044] Based on the high-definition Earth component in the virtual reality engine, the terrain of the selected sea area is dynamically modeled according to the selected longitude and latitude. The sea area environmental data is then processed and accessed through 3D interactive visualization to obtain specific sea area environmental data (temperature, salinity, depth, bottom sediments, topography, ocean currents, wind speed, eddies, ocean fronts, etc.). According to the input parameters of the acoustic field calculation model, calculation files are generated, including, for example, env water environment files (.evn), terrain elevation files (.flp), seabed shape files (.bty), and bottom reflection coefficient files (.brc). The generated files are used as input parameters to run the corresponding modeling tool (e.g., available from https: / / oalib-acoustics.org / ), generating result files, including, for example, calculation process files (.prt), sound ray files (.ray), normal mode files (.mod), sound pressure matrix files (.arr), and sound field image files (.shd). These files serve as the acoustic field characteristic information of the target sound source under the influence of the marine environment in the selected sea area, and are used to simulate the sonar signal.

[0045] Figure 3 A schematic diagram of the calculation process of fitting target features by the target feature extraction module according to an embodiment of the present application is shown.

[0046] In order to simulate the sonar signals of different targets (such as different ships), refer to Figure 3 According to an embodiment of the present invention, the target feature extraction module fits the target sound source features based on the target-related information provided by the user, which is then used to superimpose the target sound source features with the simulated signal under the dual effects of the ocean environment sound field model to achieve a comprehensive simulation of the target sound source signal propagation effect.

[0047] In sonar signal simulation, sonar targets are primarily ships, whose physical shape can be approximated as a cylinder. Ship sound source information primarily consists of mechanical noise, propeller beats, and water cavitation. Signal simulation typically employs broadband continuous spectra and narrowband characteristic line spectra. The broadband continuous spectrum is typically related to the ship's tonnage, speed, number of propellers, and rotational speed. The narrowband characteristic line spectra are inherent, invariant characteristics of a specific ship, related to its unique mechanical structure, vibration patterns, and hull material. Random simulation is typically employed. By constructing a ship feature database, the influencing factors are compared and analyzed with actual collected ship noise, and a logistic regression prediction model is employed to fit (predict) the target characteristics.

[0048] The target sound source characteristics are given by the logistic regression prediction model. The logistic regression prediction model takes the target physical shape, target tonnage, target speed, target number of propellers, target propeller speed, etc. (collectively referred to as target form) as input parameters X m, the collected ship noise data in the corresponding target form is subjected to feature information extraction, and the feature information extracted from the ship noise data includes parameters such as frequency, frequency band, bandwidth, amplitude, time-frequency characteristics and / or voiceprint spectrogram data as input parameters X m The corresponding standard output parameters in,

[0049] m: The number of attributes contained in a sample in the sample set

[0050] n: the number of samples in the sample set

[0051] i: Parameter used to traverse samples in sample set calculation.

[0052] Optionally, create a record target format (X m ) and the characteristic information of the target noise data (Y m ) database, allowing users to input or select the target form to be simulated, and obtain the characteristic information of the corresponding target noise data from the database for the training of the logistic regression prediction model and the subsequent simulation of the sonar signal.

[0053] Continue to read Figure 3 , set the iterative error value e of the learning operation, the iterative step size a of the learning operation, the prediction function h, the propagation loss function G, the randomly initialized weight vectors W0 and b0, and perform iterative calculation according to the logistic regression prediction model. t represents the number of iterations, and the initial value is 0.

[0054] Step 1: Calculate the predicted value of the current iteration round y^=W T X+b:

[0055]

[0056] Where X is the input parameter of the sample, and y^ is the predicted value of the target feature.

[0057] Step 2: Construct prediction function:

[0058]

[0059] Step 3: Calculate the loss function:

[0060]

[0061] Step 4: Calculate the loss function gradient descent weight parameter search direction:

[0062]

[0063] Step 5: Update feature vector weights: W t+1 =W t +a·Dt

[0064]

[0065] Step 6: Calculate whether the current loss function deviation is less than the set error value e. The loss function deviation is the difference between the sample parameter Y m and the loss function obtained from the current iteration (|Y m -G t |<e). The closer the difference is to zero, the closer the loss function is to the sample standard output, and thus the current parameters w and b can achieve the machine learning effect.

[0066] Repeat Step 1 until the condition in Step 6 is judged to be true.

[0067] Through the above iterative calculation process, W and b are obtained, and then the predicted value Y of the target sound source feature is obtained (y^ = W T X + b).

[0068] The target feature (Y, such as voiceprint spectrogram data) obtained according to the embodiments of the present application is an expression of the target inherent feature and is independent of the relative azimuth / distance from the target to the observation position.

[0069] Therefore, in the application of sonar signal simulation, using the above-trained logistic regression prediction model, the target sound source feature (Y) is obtained according to the set target form (target physical shape, target tonnage, target speed, number of target propellers, target propeller speed, etc.). By setting the target position and the observation point (the position where the sonar is located), the sea area where the target sound signal propagates is obtained, and the ocean environmental sound field model of the selected sea area is obtained. The target sound ray arrival effect result is calculated at the observation point position, and signal superposition calculation is performed, thereby forming the target sound signal simulation process at the observation point and obtaining the simulated sonar signal data. The sound ray arrival effect result is the sound signal result affected by the sound propagation effect of the sound source information received at the observation point, that is, the sound signal receiving position.

[0070] Optionally, to make the sound field propagation effect in the selected sea area closer to the actual effect, the sound propagation effect field data is displayed and marked by means of 3D engine visualization, and the sound pressure field data file of the selected sea area is generated. Combining the target sound source feature, the simulated data representing the sonar signal generated by the target and received by the sonar is generated. The sound pressure field data file is a file that distributes the change law of the sound pressure value in the selected area calculated according to the sound field propagation model; the target sound source feature is the basic attribute of the sound source, mainly the parameters in the frequency domain. The simulated data of the sonar signal is the receiver sound signal data generated by simulating the signal according to the distribution of the attribute parameters in the sound source feature based on the determined positions of the sound source and the sonar receiver according to the sound pressure field data file.

[0071] Figure 4 A schematic diagram showing parallel computing of simulated sonar signals by a computing platform according to an embodiment of the present application is shown.

[0072] To improve the versatility of sonar signal simulation, a configurable open interface design is adopted for different sonar devices. The number of receivers, receiver model, receiver distribution location, array element spacing, and signal processing parameters (collectively referred to as sonar configuration information) are used as interface parameters, allowing users to customize the simulation of various sonar devices.

[0073] The target sound source characteristics and the ocean environment sound field model are also provided to the signal processing platform to simulate general sonar signals. Because the signal simulation process is computationally intensive and requires parallel, synchronous, and near-real-time processing capabilities, a GPU-based signal processing platform is selected, consisting of, for example, four GPU signal processing modules and a CPU integrated control computing chip.

[0074] The signal processing platform is implemented by programming on the GPU using the CUDA computing architecture. The CUDA architecture consists of grids, thread blocks, and threads. Threads are the basic unit of parallel program execution. The GPU provides several (e.g., 2-3) grids, each containing several (e.g., 65535) thread blocks, and each thread block containing several (e.g., 512) threads.

[0075] Reference Figure 4 The GPU-based signal processing platform generates data representing the analog signal received at the sonar location after the target sound source propagates through the selected sea area, and matches the sonar equipment signal processing process to form beam domain data. After performing FFT transform on the N-channel array element domain signal data, M-channel beamforming data is formed. The operation of performing FFT transform on the N-channel array element domain signal data is implemented by a block (thread block) of the GPU. The processing of each beamforming channel is implemented by a block (thread block) of the GPU (correspondingly, M blocks (thread blocks) are used, and these M blocks belong to the same grid within the GPU). Each block provides L threads (threads), and each thread (thread) calculates the phase shift value of each array element signal in L frequency points, ultimately forming the LFFT transform of the M-channel beamforming data. The LFFT transform operation of the M-channel beamforming data is implemented by a block (computation block) of the GPU.

[0076] The operating environment / client provided by the embodiments of the present application includes, for example, a computer, a server, or other information processing device. These devices include, for example, a memory, one or more processors, one or more presentation components, I / O components, and a power supply that are directly or indirectly coupled to a bus. The bus may represent one or more types of buses (such as an address bus, a data bus, or a combination thereof).

[0077] Although preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they are aware of the basic inventive concepts. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of this application. Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if such changes and modifications fall within the scope of the claims of this application and their equivalents, then this application is intended to include such changes and modifications.

Claims

1. General sonar signal simulation method, including: Get the form X of the target to be simulated, and use the logistic regression prediction model to obtain the target sound source feature Y corresponding to the form of the target, where W is used T The result of X+b is used as the prediction of the target sound source feature Y, and W and b are the weight vectors of the logistic regression prediction model; Obtaining the positions of the target and the sonar that collects the target in earth coordinates; obtaining the sea area where the target and the sonar are located and sea area environment data of the sea area according to the positions of the earth coordinates; generating a sea area environment sound field model of the sea area according to the sea area environment data, wherein the sea area is divided into one or more sub-sea areas according to the spatial distribution of the sea area environment data, and a corresponding sound field model is matched for each sub-sea area, and the sound field models of each sub-sea area of ​​the sea area together constitute the sound field model of the sea area; The configuration information of the sonar is obtained, and the propagation effect of the target sound in the sea area is calculated based on the sound field model of the sea area and the characteristics of the target sound source. In addition, data of an analog signal representing the target sound received by the sonar is calculated in combination with the configuration information of the sonar.

2. The method according to claim 1, wherein the logistic regression prediction model is obtained by machine learning using a sample data set. The sample data set includes n samples, where n is a positive integer; The target is a ship, and the form X of the target in the sample includes the target physical shape, target tonnage, target speed, target number of propellers and / or target propeller speed. The target sound source feature Y includes the frequency, frequency band, bandwidth, amplitude, time-frequency characteristics and / or soundprint spectrogram data of the target sound.

3. The method according to claim 2, wherein performing machine learning by utilizing a sample data set comprises: Set the machine learning iteration error value e, machine learning iteration step a, prediction function h, propagation loss function G, initialization weight vector W0, b0, t represents the number of iterations; Step 1: Calculate the predicted value of the current iteration round y^=W T X+b: Where y^ is the predicted value of the target feature, n is the number of samples in the sample dataset, and m is the number of attributes in X; Step 2: Construct prediction function: Step 3: Calculate the loss function: Step 4: Calculate the loss function gradient descent weight parameter search direction: Step 5: Update feature vector weights: W t+1 =W t +a·D t , Step 6: Calculate whether the current loss function deviation is less than the iteration error value e; Repeat steps 1 to 6 until the condition in step 6 is true, and obtain the weight vectors W and b; And according to the acquired form X of the target to be simulated, and the weight vector W and the weight vector b of the trained logistic regression prediction model, the target sound source feature Y corresponding to the form of the target is predicted.

4. The method according to claim 3, wherein the sea area environmental data includes temperature, salinity, depth, bottom texture, topography, ocean current, wind speed, eddy current and / or ocean front information of the sea area; and wherein the sea area is divided into one or more sub-sea areas according to the spatial distribution of the sea area environmental data, comprising: Constructing a sea area grid according to the spatial position of the sea area, each sea area grid is an area enclosed by a specified range of longitude and latitude; Calculate the sound speed distribution assimilation value and / or the seabed elevation data change rate assimilation value for each sea area grid; Compare one or more assimilation values ​​of adjacent sea area grids, and divide the sea area into one or more sub-sea areas according to the change in the assimilation values ​​of the adjacent sea area grids, wherein in each sub-sea area, the change in the assimilation value of any adjacent sea area grid of the sub-sea area is less than a set threshold, and the change in the assimilation value of adjacent sea area grids in different sub-sea areas is not less than or greater than the set threshold.

5. The method according to claim 4, wherein matching a corresponding sound field model for each sub-sea area comprises: According to the distribution of the sea area environmental data of each sub-sea area, the corresponding sound field model is selected to construct the sound field of the sub-sea area. The hierarchical analysis method is used to determine the sound field model corresponding to the sub-sea area, including: (1) Determine the acoustic field model evaluation system and establish a hierarchical correspondence between sub-sea area environmental data and acoustic field models; (2) Construct a judgment matrix to determine the impact scale of different attributes of marine environmental data; (3) Determine the weights corresponding to different attributes of marine environmental data; (4) For each sub-sea area, the scores of each sound field model are calculated according to the judgment matrix and the weights, and the sound field model with the highest score is determined as the sound field model to be applied in the sub-sea area.

6. The method according to any one of claims 1 to 5, wherein The sonar configuration information includes the number of receivers of the sonar device, the receiver model, the distribution position of the receivers and / or the array element spacing.

7. The method according to claim 6, wherein Utilizing a GPU-based signal processing platform to calculate data representing analog signals of target sounds received by the sonar; The method further comprises: The GPU-based signal processing platform is used to simulate the sonar signal processing process to form beam domain data, wherein The operation of performing FFT transformation on N array element domain signal data is implemented by a thread block of the GPU, where N is a positive integer; Each of the M-way beamforming processes is performed by a thread block on the GPU. Each thread block for beamforming processing provides L threads, and each thread calculates the phase shift value of each array element signal at L frequency points. as well as The LFFT transform operation of the M-way beamforming data is implemented by one thread block of the GPU.

8. The method according to claim 7, wherein The M thread blocks processing each path of the M-way beamforming belong to the same grid.

9. A computing device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method according to any one of claims 1 to 8 when executing the program.

Images