A method for enhancing the line spectrum of moving targets based on elite selection genetic algorithm
By using an elite selection genetic algorithm to autonomously search for the optimal target motion parameters and employing the Doppler-warping transformation to concentrate the dispersed line spectrum energy near the original frequency, the problem of the inability of moving target line spectrum energy to accumulate over a long period of time is solved, thus improving the detection and recognition accuracy.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- THE 715TH RES INST OF CHINA SHIPBUILDING IND CORP
- Filing Date
- 2026-04-29
- Publication Date
- 2026-06-26
Smart Images
Figure CN122286445A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of underwater acoustic signal processing technology, specifically to a method for enhancing the line spectrum of moving targets based on an elite selection genetic algorithm. Background Technology
[0002] The radiated noise spectrum of a moving sound source typically consists of several discrete single-frequency line spectra and a broadband continuous spectrum. The line spectra are mainly concentrated in the low frequencies, possessing advantages such as concentrated energy, good stability, and long propagation distance, making them important characteristics for underwater target detection and identification. However, when the sound source is in motion while the receiving point is stationary, the received signal will experience a Doppler frequency shift. The originally stable line spectrum signal will become a non-stationary signal with frequency varying over time, resulting in dispersed line spectrum energy that cannot accumulate over long periods, severely reducing the target detection probability and identification accuracy.
[0003] Warping transform is a unitary equivalent transform based on the dispersion characteristics of normal modes. It utilizes the instantaneous phase expression of normal modes and, through resampling in the time or frequency domain according to a specific mapping relationship, can convert complex non-stationary signals into simple quasi-single-frequency signals, thereby achieving the separation of different normal modes. Doppler-warping transform is a signal processing method that applies warping transform to the classical acoustic Doppler problem, achieving linearization of the Doppler frequency shift. However, the effective implementation of Doppler-warping transform depends on the accurate given target motion parameters, including the target's true velocity, the closest distance between the target trajectory and the receiving point, and the time when the target reaches the closest point. Only when these parameters are accurately matched can the instantaneous frequency of the received signal line spectrum be linearized, and the signal spectral energy be refocused near the original target frequency. When the input target motion parameters deviate from their true values, the transformed instantaneous frequency remains nonlinear, and the spectral energy remains dispersed. The greater the deviation, the more dispersed the spectral energy, and the worse the line spectrum enhancement effect.
[0004] Therefore, how to autonomously and accurately search for target motion parameters to optimize the configuration of Doppler-warping transform parameters, thereby solving the problem that the energy of the moving target line spectrum cannot be accumulated over a long period of time, has become a technical problem that urgently needs to be solved in the field of underwater acoustic signal processing. Summary of the Invention
[0005] To address the shortcomings of existing technologies, the present invention aims to provide a method for enhancing the line spectrum of moving targets based on an elite selection genetic algorithm.
[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for enhancing the line spectrum of a moving target based on an elite selection genetic algorithm, characterized by comprising the following steps: Acquire the radiation signal of a moving target, wherein the radiation signal of the moving target generates a Doppler frequency shift due to the target's motion, and the line spectrum signal is a non-stationary signal whose frequency changes with time; Based on the elite selection genetic algorithm, the motion parameters of the moving target are estimated by using the entropy of the target signal spectrum function after Doppler-warping transformation as the cost function, and the optimal target motion parameters are obtained. By using the optimal target motion parameters to perform a Doppler-warping transformation on the radiation signal of the moving target, the line spectrum energy dispersed due to the Doppler effect is refocused near the original frequency, thereby achieving autonomous enhancement of the moving target line spectrum.
[0007] In some embodiments, the motion parameters include the target's speed, the closest distance between the target trajectory and the receiving point, and the time when the target reaches the closest point.
[0008] In some embodiments, the estimation of motion parameters of the moving target based on an elite selection genetic algorithm specifically includes: Initialize the population by randomly generating a set of initial solutions as the first generation population, where each individual represents a set of target motion parameters; For the different motion parameters represented by each individual in the population, the Doppler-warping transformation is performed on the radiation signal of the moving target, and the entropy of the spectrum function of the transformed target signal is calculated as the cost function value. The elite selection genetic algorithm is used to select, crossbreed, and mutate the population, retaining elite individuals and iteratively updating the population. Determine whether the cost function value of the current best individual meets the preset termination condition. If it does, the algorithm terminates and outputs the motion parameters corresponding to the current best individual as the optimal target motion parameters.
[0009] In some embodiments, the cost function is the entropy of the target signal spectrum function after Doppler-warping transformation. By comparing the entropy values of the target signal spectrum function under different motion parameters, the motion parameters corresponding to the minimum entropy value are the optimal target motion parameters.
[0010] In some embodiments, the Doppler-warping transform is a unitary equivalent transform based on the dispersion characteristics of normal modes. By resampling in the time or frequency domain according to a specific mapping relationship, the complex non-stationary signal generated by the Doppler effect is converted into a quasi-single-frequency signal, thereby linearizing the instantaneous frequency of the received signal line spectrum.
[0011] In some embodiments, the parameters of the elite selection genetic algorithm include: initial population size, crossover rate, selection rate, mutation rate, and iteration termination threshold.
[0012] In some embodiments, the preset termination condition is to stop iteration when the cost function value is less than a set threshold, and the set threshold is set according to actual processing requirements.
[0013] In some embodiments, due to the randomness of the optimization results, the average of multiple optimization results is used as the final optimal target motion parameters.
[0014] In some embodiments, the method is applied to data backtracking or data replay processing for underwater moving target detection, and is used to detect and identify the line spectrum of moving targets in scenarios with low real-time requirements.
[0015] Compared with the prior art, the beneficial effects of the present invention are: (1) Long-term accumulation and autonomous enhancement of line spectrum energy of moving targets are realized. Based on the Doppler-warping transform, this invention uses an elite selection genetic algorithm to autonomously optimize and estimate the target motion parameters, and reconcentrates the line spectrum energy dispersed by the Doppler effect near the original frequency, effectively solving the problem that the line spectrum energy of moving targets cannot be accumulated over a long period of time, and significantly improving the detection probability and detection accuracy of the line spectrum.
[0016] (2) Autonomous search of Doppler-warping transform parameters is realized. This invention uses the entropy of the target signal spectrum function after Doppler-warping transform as the cost function and automatically searches for the optimal target motion parameters through an elite selection genetic algorithm. No manual intervention or prior precise parameter input is required, which overcomes the shortcomings of traditional Doppler-warping transform that depends on real parameters and whose enhancement effect drops sharply when the parameters deviate.
[0017] (3) Effective linearization and frequency recovery of Doppler frequency shift signal are achieved. By performing Doppler-warping transformation on the original signal with optimal motion parameters, the complex non-stationary Doppler signal generated by the target motion can be converted into a quasi-single frequency signal, which linearizes the instantaneous frequency of the received signal line spectrum, thereby accurately recovering the original frequency characteristics of the target radiation signal and providing reliable support for target identification and tracking.
[0018] (4) It has good engineering practical value. The principle of the method of the present invention is clear and the implementation method is well-defined. It has been verified by computer simulation and actual sea trial data processing. It can accurately estimate the target motion state and effectively enhance the target line spectrum. It can be widely used in underwater moving target detection data backtracking system and data review system with low real-time requirements. It has strong engineering applicability and promotion value.
[0019] Details of one or more embodiments of this application are set forth in the following drawings and description to make other features, objects and advantages of this application more readily apparent. The embodiments of this application will provide a detailed description and understanding of the application. Attached Figure Description
[0020] Figure 1 This is a flowchart of the method of the present invention; Figure 2 Here is a flowchart of the genetic algorithm based on elite selection; Figure 3 This is a schematic diagram of a sound source motion model; Figure 4 This is a comparison image before and after the Doppler-warping transformation; Figure 5 The result of processing simulation data using the method described in this invention; Figure 6 This is a graph showing the signal history of an actual moving target. Figure 7 The result is obtained by processing actual data using the method described in this invention. Detailed Implementation
[0021] 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.
[0022] Example 1: Overall Flow of a Motion Target Line Spectrum Enhancement Method Based on Elite Selection Genetic Algorithm Please see Figure 1 This embodiment provides an overall implementation process for a motion target line spectrum enhancement method based on an elite selection genetic algorithm, including the following steps: Step S1: Acquire the radiated signal of the moving target. The radiated signal of the moving target experiences a Doppler frequency shift due to the target's motion, and the line spectrum signal is a non-stationary signal whose frequency varies with time. Specifically, the radiated noise signal of the moving target is acquired by a hydrophone array deployed at the underwater receiving point. After pre-amplification, anti-aliasing filtering, and analog-to-digital conversion, the signal is obtained as a discrete digital signal sequence. The radiated noise spectrum of a moving sound source typically consists of several discrete single-frequency line spectra and a broadband continuous spectrum. The line spectrum is mainly concentrated in the low frequency range, possessing advantages such as concentrated energy, good stability, and long propagation distance. When the sound source is in motion and the receiving point is stationary, the received signal will experience a Doppler frequency shift, and the originally stable line spectrum signal will become a non-stationary signal whose frequency varies with time, resulting in dispersed line spectrum energy and an inability to accumulate over a long period.
[0023] Step S2: Based on the elite selection genetic algorithm, the motion parameters of the moving target are estimated using the entropy of the target signal spectrum function after the Doppler-warping transform as the cost function, to obtain the optimal target motion parameters. These motion parameters include the target's velocity v, the closest distance r0 between the target trajectory and the receiving point, and the time t0 when the target reaches the closest point. The Doppler-warping transform is a unitary equivalent transform based on the dispersion characteristics of normal modes. By resampling in the time or frequency domain according to a specific mapping relationship, the complex non-stationary signal generated by the Doppler effect is converted into a quasi-single-frequency signal, thus linearizing the instantaneous frequency of the received signal line spectrum.
[0024] Step S3: Using the optimal target motion parameters, perform a Doppler-warping transformation on the moving target radiation signal to refocus the line spectrum energy dispersed by the Doppler effect near the original frequency, thereby achieving autonomous enhancement of the moving target line spectrum.
[0025] In this embodiment, the Doppler-warping transform is implemented as follows: based on the dispersion characteristics of normal modes, the instantaneous phase expression of normal modes is used to convert complex non-stationary signals into simple quasi-single-frequency signals by resampling in the time or frequency domain according to a specific mapping relationship. The warping transform operator is the inverse function of the instantaneous phase of the signal, calculated from the target motion parameters. An energy conservation factor is introduced during the transform process to ensure that the signal energy remains unchanged before and after the transform. When the input target motion parameters accurately match the true values, the instantaneous frequency of the transformed received signal line spectrum is linearized, and the spectral energy is reconcentrated near the original target frequency; when the input parameters deviate from the true values, the transformed instantaneous frequency is still nonlinear, and the spectral energy is still dispersed, and the greater the deviation, the more dispersed the spectral energy.
[0026] Example 2: Optimal Motion Parameter Search Method Based on Elite Selection Genetic Algorithm Please see Figure 2 This embodiment details the specific process of searching for the optimal motion parameters of a moving target based on an elite selection genetic algorithm, including the following sub-steps: Step S201: Initialize the population. Based on the known medium sound speed c, set the search range for the target's velocity v, the closest distance r0 between the target trajectory and the receiving hydrophone, and the time t0 when the sound source reaches that point. Simultaneously, set the upper and lower limits f of the target's line spectrum frequency. min f max Within the aforementioned search range, a set of initial solutions is randomly generated as the first generation population, with each individual representing a set of target motion parameters (v, r0, t0).
[0027] Step S202: Fitness Assessment. For each individual in the population, a Doppler-warping transform is performed on the radiated signal of the moving target, and the entropy of the spectrum function of the transformed target signal is calculated as the cost function value to determine the fitness of each individual in the population.
[0028] The cost function is the spectrum function S(f) of the target signal after Doppler-warping transformation in the frequency band [f]. min ,f max The entropy H within [ ] is expressed as follows: (Formula 1: Spectral Entropy Cost Function) Where S(f) is the transformed signal s m (t) is obtained by performing a Fourier transform and taking the modulus, i.e.: (Formula 2: Definition of Spectrum Function) s m (t) is obtained from the original signal through Doppler-warping transformation. The Doppler-warping transformation process is as follows: (Formula 3: Doppler-warping transformation relationship) Where a(τ) is the warping transform operator, s m (τ) represents the transformed signal, and the role of a(τ) is to ensure energy conservation before and after the transformation. The warping operator a(τ) is the inverse function of the instantaneous phase of the signal, which can be calculated from the target motion parameters v, r0, and t0.
[0029] For each individual in the population, a cost function is calculated for different motion parameters. By comparing the entropy values of the target signal spectrum function under different motion parameters, the motion parameters corresponding to the minimum entropy value are the optimal target motion parameters.
[0030] Step S203: Use an elite selection genetic algorithm to perform selection, crossover, and mutation operations on the population, retaining elite individuals and iteratively updating the population. Specifically, this includes: selecting a subset of individuals as parents to generate the next generation based on their fitness values (spectral entropy values); performing crossover on the selected parent individuals to generate new individuals; and mutating the new individuals to increase population diversity. Simultaneously, the best individual in the current generation (i.e., the individual with the lowest entropy value) is directly retained in the next generation to ensure that excellent genes are not lost.
[0031] Step S204: Determine whether the cost function value of the current optimal individual meets the preset termination condition. If it does, the algorithm terminates and outputs the motion parameters corresponding to the current optimal individual as the optimal target motion parameters; otherwise, return to step S202 to continue iteration.
[0032] In this embodiment, the parameters of the elite selection genetic algorithm include: initial population size N. pop Hybridization rate P c Selection rate P s Variation rate P m and the iteration termination threshold H th The preset termination condition is when the cost function value (spectral entropy) is less than a set threshold H. th The iteration stops when the threshold is set, and the threshold is set according to the actual processing requirements.
[0033] Example 3: Robust Parameter Estimation Method Based on Multiple Optimizations and Mean Taking This embodiment addresses the issue of randomness in the optimization results of genetic algorithms by proposing a robust parameter estimation method that involves multiple optimizations and averaging to improve the stability and reliability of optimal motion parameter estimation.
[0034] Since the initial population of the genetic algorithm is randomly generated, and selection, crossover, and mutation operations all involve random factors, the results of a single optimization attempt may fluctuate. To suppress the impact of this randomness on the accuracy of the final parameter estimation, this embodiment adopts the following strategy: The algorithm performs M independent elite selection genetic algorithm optimizations (M being a preset number of iterations, e.g., M=50). Each optimization uses the same search range and algorithm parameter settings, but with different random seeds, thus obtaining M sets of optimal motion parameter estimates. The average values of the target velocity v, the nearest distance r0, and the arrival time t0 are calculated for each of the M optimization results, and the average value of each parameter is used as the final optimal target motion parameters output.
[0035] In addition, the standard deviation or average deviation of the M optimization results can be calculated as an auxiliary indicator for evaluating the estimation accuracy. If the deviation of a certain parameter exceeds a preset threshold, the number of optimization attempts M can be appropriately increased or the genetic algorithm parameters can be adjusted (such as increasing the population size or reducing the mutation rate) to further improve the estimation accuracy.
[0036] Example 4: Computer Simulation Verification 4.1 Simulation Condition Settings To verify the effectiveness of the method of this invention, computer simulation was conducted. It is assumed that the target is moving in uniform linear motion in seawater, the target radiated noise frequency is f0 = 150 Hz, the closest distance between the target trajectory and the receiving point is r0 = 200 m, the arrival time of the nearest sound source is t0 = 60 s, the received signal time range is T = 0-120 s, the speed of sound in water is c = 1500 m / s, the target's velocity is v = 5 m / s, and the signal sampling rate is f. s =5kHz. Use the above parameters to construct Doppler signal simulation data.
[0037] Under the above simulation conditions, the time-frequency characteristics and amplitude spectrum comparison of the signals before and after the Doppler-warping transform are as follows: Figure 4 As shown (corresponding to the original application documents) Figure 3 ).from Figure 4 It can be seen that, under the current simulation conditions, the Doppler-warping transform can linearize the Doppler phase of a moving target. Due to the Doppler frequency shift effect, the original signal's spectrum exhibits bandwidth broadening near f0. After the Doppler-warping transform, the energy is concentrated again near the original frequency, which is beneficial for the target's line spectrum detection. However, it should be noted that when the selected target motion parameters are mismatched, the Doppler signal cannot be effectively restored.
[0038] 4.2 Simulation of Parameter Optimization Based on Elite Selection Genetic Algorithm The target motion parameters under the above simulation conditions are optimized using an elite selection genetic algorithm. The algorithm parameters are set as follows: initial population size is 100, crossover rate is 0.7, selection rate is 0.5, mutation rate is 0.01, and the number of iterations is determined by the cost function. Iteration stops when the cost function (spectral entropy) is less than a set threshold. The threshold value is set according to the actual situation; under the current simulation conditions, the threshold is set to -200.
[0039] Due to the randomness inherent in the optimization results, 50 independent optimizations were performed, and the average was taken as the final result. The results of the 50 optimizations are shown in Table 1: Table 1 Optimization results of target motion parameters (simulation data) As can be seen from Table 1, the average target motion speed of the 50 optimization results is 5.0050 m / s (true value 5 m / s), the average closest distance is 203.0836 m (true value 200 m), and the average arrival time is 60.0469 s (true value 60 s). The average values of the three parameters are very close to the true values, which verifies the accuracy of the motion parameter estimation by the method of the present invention.
[0040] The search results from the elite selection genetic algorithm are used to perform a Doppler-warping transform on the original signal. The time history plot and amplitude spectrum of the transformed signal are shown below. Figure 5 As shown (corresponding to the original application documents) Figure 4 The results show that the elite selection genetic algorithm can accurately search for the target motion parameters, complete the recovery of the Doppler signal, effectively concentrate the line spectrum energy around the original frequency of 150Hz, significantly reduce the spectral entropy, and achieve a significant line spectrum enhancement effect.
[0041] Example 5: Processing of Actual Sea Trial Data To further verify the effectiveness of the method of this invention in a real engineering environment, the moving target data actually measured during sea trials was processed. The total data length is 300s. A short-time Fourier transform was performed on this signal, revealing a significant Doppler frequency shift signal around 172Hz, which was then used as the research object. The time-frequency image of the actual moving target signal is shown below. Figure 6 As shown (corresponding to the original application documents) Figure 5 ).
[0042] The target parameters corresponding to the Doppler signal were estimated using a target parameter estimation method based on an elite selection genetic algorithm. The search range was set as follows: the discussion interval for the target velocity v was 0-10 m / s, the discussion interval for the nearest distance r0 was 0-2000 m, and the discussion interval for the arrival time t0 was 100-200 s. The parameter settings of the elite selection genetic algorithm were the same as in Example 4. The estimation results are shown in Table 2. Table 2 Optimization results of actual target motion parameters The search results based on an elite selection genetic algorithm were used to perform a Doppler-warping transform on the original signal. The time history plot and amplitude spectrum of the transformed signal are shown below. Figure 7 As shown (corresponding to the original application documents) Figure 6 The results show that using the target motion parameters obtained by the elite selection genetic algorithm can effectively restore the target signal frequency. The originally dispersed Doppler frequency shift energy is reconcentrated around the original frequency of 172Hz, and the line spectrum enhancement effect is significant, proving the effectiveness and engineering applicability of the method of this invention in actual underwater moving target detection.
[0043] Example 6: Application in a Data Backtracking and Retrospective System This embodiment illustrates a typical application scenario of the method of the present invention in data backtracking or data review processing for underwater moving target detection.
[0044] Because this invention uses an elite selection-based genetic algorithm in step S2 to search for the optimal motion parameters of the moving target, although it achieves autonomous search for the Doppler-warping transformation parameters, the search time is relatively high and the processing results cannot be fed back in a timely manner. Therefore, this method is not suitable for real-time processing systems. The main purpose of this invention is to achieve autonomous search for Doppler-warping transformation parameters, which can be applied to data backtracking systems and data replay systems with lower real-time requirements.
[0045] Specific application scenarios include: (1) Retrospective processing of underwater moving target detection data: In the post-event data processing stage, the underwater acoustic signal data containing moving targets collected in history are retrospectively analyzed. The target motion parameters are automatically estimated and the line spectrum is enhanced using the method of this invention. The line spectrum features of the target are extracted for target identification and classification.
[0046] (2) Review and processing of underwater moving target detection data: Review and analyze the detected events, reprocess the original data using the method of this invention, optimize the parameter estimation results, verify the previous detection conclusions, or provide data support for tactical assessment and training.
[0047] (3) Target motion situation estimation: By backtracking multiple batches of historical data, the parameter estimation results at multiple times are accumulated, and the motion trajectory and speed change pattern of the target are inverted, providing a basis for target motion situation analysis.
[0048] In the aforementioned application scenarios, the method of this invention can be integrated into an underwater acoustic signal processing software platform and run independently as a post-processing module. Users only need to input the original signal data and a basic search range, and the system can automatically complete parameter optimization, line spectrum enhancement, and result output without manual intervention, demonstrating significant engineering practical value.
[0049] Example 7: Alternative Optimization Algorithm Scheme This embodiment provides an alternative implementation of the parameter optimization step in the method of the present invention. In embodiments one to six, step S2 uses an elite selection genetic algorithm to search for optimal motion parameters. As an alternative, this step can be replaced by other global optimization algorithms, including but not limited to: (1) Particle Swarm Optimization (PSO): Each particle is regarded as a set of candidate motion parameters. The optimal solution is found by the flight of particles in the parameter space and the sharing of group information. The spectral entropy is used as the fitness function.
[0050] (2) Simulated Annealing Algorithm: By simulating the physical annealing process, the spectral entropy is used as the energy function to perform a random search in the parameter space. The Metropolis criterion is used to accept inferior solutions and avoid getting trapped in local optima.
[0051] (3) Differential Evolution Algorithm: Through differential mutation, crossover and selection operations, differential vectors are introduced into the population to guide the search direction, and the optimization objective is to minimize the spectral entropy.
[0052] All of the above-mentioned alternative optimization algorithms can use spectral entropy as the cost function, and iteratively search for the combination of motion parameters that minimizes spectral entropy, thereby achieving autonomous optimization of the Doppler-warping transform parameters and achieving the goal of enhancing the line spectrum of the moving target. When using different optimization algorithms, the algorithm parameters need to be adjusted according to the specific problem, but the basic process is consistent with the process described in Embodiment 2 of this invention.
[0053] Example 8: The specific implementation process of the present invention is as follows: 1) Given the speed of sound in the medium Set target speed The closest distance between the target and the receiving hydrophone And the moment when the sound source moves to that point The search range is set by defining the upper and lower limits of the target line spectrum frequency. , .
[0054] 2) Initialize the population by randomly generating an initial solution as the first generation population.
[0055] 3) Fitness assessment: Each individual in the population is evaluated to determine its fitness. The fitness is calculated using a cost function, which is the spectrum function of the signal after the Doppler-warping transform. In frequency band internal entropy .entropy The expression is as follows: in, From the transformed signal The Fourier transform is performed and the modulus is taken to obtain the result. The signal is obtained by Doppler-warping transformation of the original signal. The Doppler-warping transformation process is as follows: in, For the warping transformation operator, The transformed signal, Its function is to ensure energy conservation before and after the transformation. (warping operator) It is the inverse function of the instantaneous phase of the signal, and can be derived from the target motion parameters. , , Calculated and written as For each individual in the population, a cost function is calculated for the different motion parameters they represent.
[0056] 4) Use an elite selection genetic algorithm to search for the optimal motion parameters of the moving target.
[0057] The selection, crossover, and mutation operations are the same as in conventional genetic algorithms. Based on the fitness values of individuals, a subset of individuals are selected as parents to produce the next generation. These selected parents are then crossoverdone to generate new individuals, which are then mutated to increase population diversity. Elite retention ensures that the best individuals from the current generation are directly carried over to the next generation, preventing the loss of superior genes. A termination condition is set; if the preset termination condition is met, the algorithm stops; otherwise, it returns to step 3 to continue iteration. This invention sets a termination threshold of [insert threshold here]. And when the fitness of the best individual in a certain generation is less than When the time is reached, the algorithm terminates, and that individual is the optimal solution.
[0058] 5) Use the optimal target motion parameters to perform Doppler-warping transformation on the original signal to achieve autonomous enhancement of the moving target line spectrum.
[0059] The original signal is subjected to Doppler-warping transformation using the target optimal motion parameters obtained by the elite selection genetic algorithm. The transformation process is as follows: in, The optimal warping transformation operator can be derived from the target's optimal motion parameters. , , Calculated and written as Computer Simulation 1: Assume the target is moving at a constant velocity in a straight line in seawater, and the trajectory is as follows: Figure 3 As shown, the target radiated noise frequency is 150Hz, the closest distance is 200m, the closest arrival time of the sound source is 60s, the received signal time range is 0-120s, the speed of sound in water is 1500m / s, the target's moving speed is 5m / s, and the signal sampling rate is 5kHz. The Doppler signal constructed using the above parameters and the results after Doppler-warping transform are shown below. Figure 4 (a) Figure 4 As shown in (b). Figure 4 (c) Comparison of the amplitude spectrum of the entire signal before and after the Doppler-warping transform. It can be seen that under the current simulation conditions, the Doppler-warping transform can linearize the Doppler phase of the moving target. Due to the Doppler frequency shift effect, Figure 4 (c) The spectrum of the original signal shows a bandwidth broadening phenomenon in the vicinity. After Doppler-warping transformation, the energy is concentrated near the original frequency, which is beneficial to the line spectrum detection of the target. However, it should be noted that when the selected target's motion speed is mismatched, the Doppler signal cannot be effectively restored.
[0060] Computer Simulation 2: The following section uses an elite selection genetic algorithm to optimize the target motion parameters based on the free-field simulation results from Computer Simulation 1. The elite selection genetic algorithm sets the initial population size to 100, the crossover rate to 0.7, the selection rate to 0.5, and the mutation rate to 0.01. The number of iterations is determined by the cost function; iteration stops when the cost function is less than a set threshold. The threshold is set according to the actual situation; under the current simulation conditions, the threshold is set to -200. Due to the randomness of the optimization results, the average of multiple optimization results is used as the final result. The results of 50 optimizations are shown in Table 1. The search results from the elite selection genetic algorithm are used to perform a Doppler-warping transform on the original signal. The time-frequency plot and amplitude spectrum of the transformed signal are shown below. Figure 5 As shown in the results, the elite selection genetic algorithm can accurately search for the target motion parameters and complete the recovery of the Doppler signal.
[0061] Experimental data processing: The total data length is 300s. A short-time Fourier transform was performed on this signal. A Doppler signal exists near 172Hz, and this signal is used as the research object. Its time-frequency image is as follows. Figure 6 As shown in Table 2, the target parameters corresponding to the Doppler signal were estimated using an elite selection genetic algorithm-based target parameter estimation method. The estimation results are shown in Table 2. Furthermore, the Doppler-warping transform was performed on the original signal using the search results from the elite selection genetic algorithm. The time-frequency plot and amplitude spectrum of the transformed signal are shown in Table 2. Figure 7 As shown in the results, the target motion parameters obtained using an elite selection-based genetic algorithm can effectively achieve target signal frequency recovery.
[0062] The simulation results above demonstrate that this invention proposes an autonomous enhancement method for the line spectrum of moving targets based on an elite selection genetic algorithm, addressing the problem of underwater moving target detection. This method, based on Doppler-warping transform, utilizes an elite selection genetic algorithm to estimate the motion parameters of the moving target, thereby achieving frequency recovery of the target's radiated signal. Processing data from moving targets actually measured during sea trials shows that the proposed method effectively estimates the target's motion state and enhances the target's line spectrum.
[0063] The above embodiments merely illustrate several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.
[0064] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.
Claims
1. A method for enhancing the line spectrum of moving targets based on an elite selection genetic algorithm, characterized in that: Includes the following steps: Acquire the radiation signal of a moving target, wherein the radiation signal of the moving target generates a Doppler frequency shift due to the target's motion, and the line spectrum signal is a non-stationary signal whose frequency changes with time; Based on the elite selection genetic algorithm, the motion parameters of the moving target are estimated by using the entropy of the target signal spectrum function after Doppler-warping transformation as the cost function, and the optimal target motion parameters are obtained. By using the optimal target motion parameters to perform a Doppler-warping transformation on the radiation signal of the moving target, the line spectrum energy dispersed due to the Doppler effect is refocused near the original frequency, thereby achieving autonomous enhancement of the moving target line spectrum.
2. The method for enhancing the line spectrum of moving targets based on an elite selection genetic algorithm according to claim 1, characterized in that, The motion parameters include the target's speed, the closest distance between the target's trajectory and the receiving point, and the time when the target reaches the closest point.
3. The method for enhancing the line spectrum of moving targets based on an elite selection genetic algorithm according to claim 1, characterized in that, The estimation of motion parameters of a moving target based on an elite selection genetic algorithm specifically includes: Initialize the population by randomly generating a set of initial solutions as the first generation population, where each individual represents a set of target motion parameters; For the different motion parameters represented by each individual in the population, the Doppler-warping transformation is performed on the radiation signal of the moving target, and the entropy of the spectrum function of the transformed target signal is calculated as the cost function value. The elite selection genetic algorithm is used to select, crossbreed, and mutate the population, retaining elite individuals and iteratively updating the population. Determine whether the cost function value of the current best individual meets the preset termination condition. If it does, the algorithm terminates and outputs the motion parameters corresponding to the current best individual as the optimal target motion parameters.
4. The method for enhancing the line spectrum of moving targets based on an elite selection genetic algorithm according to claim 3, characterized in that, The cost function is the entropy of the target signal spectrum function after the Doppler-warping transform. By comparing the entropy values of the target signal spectrum function under different motion parameters, the motion parameters corresponding to the minimum entropy value are the optimal target motion parameters.
5. The method for enhancing the line spectrum of a moving target based on an elite selection genetic algorithm according to claim 1 or 3, characterized in that, The Doppler-warping transform is a unitary equivalent transform based on the dispersion characteristics of normal modes. By resampling in the time or frequency domain according to a specific mapping relationship, the complex non-stationary signal generated by the Doppler effect is converted into a quasi-single-frequency signal, thereby linearizing the instantaneous frequency of the received signal line spectrum.
6. The method for enhancing the line spectrum of moving targets based on an elite selection genetic algorithm according to claim 3, characterized in that, The parameters of the elite selection genetic algorithm include: initial population size, crossover rate, selection rate, mutation rate, and iteration termination threshold.
7. The method for enhancing the line spectrum of moving targets based on an elite selection genetic algorithm according to claim 6, characterized in that, The preset termination condition is to stop iteration when the cost function value is less than a set threshold, and the set threshold is set according to the actual processing requirements.
8. The method for enhancing the line spectrum of moving targets based on an elite selection genetic algorithm according to claim 3, characterized in that, Because the optimization results are random, the average of multiple optimization results is used as the final optimal target motion parameters.
9. The method for enhancing the line spectrum of moving targets based on an elite selection genetic algorithm according to claim 1, characterized in that, The method is applied to data backtracking or data review in underwater moving target detection, and is used to detect and identify moving target line spectra in scenarios with low real-time requirements.