Adaptive underwater acoustic positioning correction method and system based on swarm intelligence

Abstract

The application relates to the technical field of underwater acoustic positioning, in particular to a self-adaptive underwater acoustic positioning correction method and system based on swarm intelligence, which comprises the following steps: reconstructing original signals of mutual detection of underwater acoustic reference nodes in a high-dimensional phase space, and extracting an arrival time characteristic vector from a phase domain trajectory; based on multi-target particle swarm optimization, performing multi-target optimization on a local neighborhood geometric structure of the characteristic vector to obtain optimal local linear embedding parameters; according to the parameters, globally splicing the characteristic vector manifold to obtain a time delay space manifold; performing variational mode decomposition on the time delay space manifold to obtain a dynamic deviation residual; according to a projection position of a target emission signal on the manifold, fusing the dynamic deviation residual to correct time delay distortion, and obtaining a self-adaptive time delay correction amount; and analyzing the correction amount intersection positioning to obtain a space position of a target to be positioned; the application can improve the correction efficiency of self-adaptive underwater acoustic positioning.

Description

Technical Field

[0001] This invention relates to the field of underwater acoustic positioning technology, and in particular to an adaptive underwater acoustic positioning correction method and system based on swarm intelligence. Background Technology

[0002] Underwater acoustic positioning technology is the core of underwater target detection and positioning, and is widely used in marine engineering fields such as marine resource exploration and underwater vehicle navigation. Its positioning accuracy and correction efficiency are directly related to the safety of underwater operations. The complex underwater environment can easily cause time delay distortion in the original signals collected by the reference node, and the accuracy of the time delay parameter is the key to positioning. Therefore, the adaptive correction of the time delay parameter has become a research focus in this field. Swarm intelligence algorithms, with their global optimization and adaptive characteristics, have been gradually applied to signal processing and parameter optimization. Particle swarm optimization has also been tried to be used for underwater acoustic positioning parameter optimization, providing a new approach to optimizing positioning correction parameters.

[0003] Existing technologies struggle to accurately mine the phase domain trajectory features of high-dimensional signals when extracting features from raw signals acquired by underwater acoustic reference nodes. They fail to effectively extract feature vectors that comprehensively characterize signal arrival times, resulting in insufficient validity of the basic data for subsequent parameter optimization. When using swarm intelligence algorithms such as particle swarm optimization for positioning correction parameters, they only optimize parameters for a single target, without considering the multi-dimensional requirements of underwater acoustic positioning to perform multi-target optimization of the local neighborhood geometry of the feature vectors. This makes it difficult for the optimized parameters to adapt to the global correction requirements of underwater acoustic positioning, affecting the overall accuracy and efficiency of positioning correction. Therefore, how to improve the correction efficiency of adaptive underwater acoustic positioning has become an urgent problem to be solved. Summary of the Invention

[0004] This invention provides an adaptive underwater acoustic localization correction method and system based on swarm intelligence to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides an adaptive underwater acoustic localization correction method based on swarm intelligence, comprising: S1. The original signals collected by the underwater acoustic reference nodes during mutual detection are reconstructed in high-dimensional phase space, and the arrival time feature vector of the original signals is extracted from the reconstructed phase domain trajectory. S2. Based on the multi-objective particle swarm optimization strategy, the local neighborhood geometry of the arrival time feature vector is optimized in multiple objectives to obtain the optimal local linear embedding parameters of the arrival time feature vector. S3. Based on the optimal local linear embedding parameters, the arrival time feature vectors are globally stitched together to obtain the time delay spatial manifold of the underwater acoustic reference node; S4. Perform variational mode decomposition on the time-delayed spatial manifold to obtain the dynamic deviation residuals of the underwater acoustic reference node; S5. Based on the projection position of the transmitted signal of the target to be located on the time delay space manifold, and by fusing the dynamic deviation residual to perform time delay distortion correction, the adaptive time delay correction amount of the target to be located is obtained. S6. Perform intersection positioning analysis on the adaptive time delay correction to obtain the spatial position of the target to be located.

[0006] In a preferred embodiment, the extraction of the arrival time feature vector of the original signal includes: The phase space of the original signals acquired by the underwater acoustic reference nodes during mutual detection is reconstructed to obtain the high-dimensional phase space of the original signals; Based on the embedding dimension and delay time determined by traversal optimization, the embedding phase point mapping is performed on the high-dimensional phase space to obtain the phase point set of the original signal; Neighborhood trajectory tracing is performed on the set of phase points to obtain the phase domain trajectory of the original signal; Local geometric feature analysis is performed on the phase domain trajectory, and the arrival time feature vector of the original signal is extracted from the analysis results.

[0007] In a preferred embodiment, obtaining the optimal local linear embedding parameters of the arrival time feature vector includes: The arrival time feature vector is partitioned into its neighborhood to obtain the local geometric neighborhood of the arrival time feature vector; Based on a multi-objective particle swarm optimization strategy, the quality of neighborhood reconstruction is evaluated for local geometric neighborhoods, and the fitness evaluation index of local geometric neighborhoods is obtained. Based on the fitness evaluation index, the local geometric neighborhood is iteratively optimized and selected to obtain high-quality embedding candidates for the local geometric neighborhood; Topological consistency verification is performed on high-quality embedding candidates to obtain stable embedding results in local geometric neighborhoods; Based on the stable embedding results, a global optimal matching is performed to obtain the optimal local linear embedding parameters for the arrival time feature vector.

[0008] In a preferred embodiment, obtaining the time-delay spatial manifold of the underwater acoustic reference node includes: Based on the optimal local linear embedding parameters, local linear embedding is performed on the arrival time feature vector to obtain the local low-dimensional embedding coordinates of the arrival time feature vector; Based on the optimal local linear embedding parameters, the local low-dimensional embedding coordinates are globally concatenated to obtain a global permutation matrix of arrival time feature vectors; Perform eigenspectral decomposition on the global permutation matrix to obtain the low-dimensional eigencoordinates of the global permutation matrix; Based on the low-dimensional intrinsic coordinates, the time delay observation data of the underwater acoustic reference node are calibrated using manifold coordinates to obtain the time delay spatial manifold of the underwater acoustic reference node.

[0009] In a preferred embodiment, obtaining the dynamic deviation residual of the underwater acoustic reference node includes: Perform decomposition parameter initialization on the time-delay space manifold to obtain the number of modal decompositions and the penalty factor of the time-delay space manifold; Based on the number of modal decompositions and the penalty factor, the intrinsic modal components of the time-delay spatial manifold are obtained by iterative separation of the intrinsic modes. Residual component identification is performed on the intrinsic modal components to obtain the trend term components of the time-delay spatial manifold; Trend term deviation analysis is performed on the trend term components to obtain the dynamic deviation residuals of the underwater acoustic reference node.

[0010] In a preferred embodiment, the trend term components of obtaining the time-delayed spatial manifold include: The intrinsic modal components are evaluated for trend characteristics to obtain trend evaluation indices for the intrinsic modal components. The trend assessment indicators are subjected to trend mode screening to obtain the trend mode components of the trend assessment indicators; By integrating and reconstructing the trend modal components, the trend term components of the time-delayed spatial manifold are obtained.

[0011] In a preferred embodiment, obtaining the adaptive time delay correction amount for the target to be located includes: Perform manifold projection positioning on the transmitted signal of the target to be located to obtain the geometric projection position of the transmitted signal on the time delay space manifold; Based on the geometric projection position, the initial time delay of the transmitted signal is extracted to obtain the initial time delay estimate of the target to be located; Based on the dynamic deviation residual, the initial time delay estimate is corrected by deviation compensation to obtain the compensated time delay of the target to be located. The time delay is corrected for distortion after compensation to obtain the adaptive time delay correction for the target to be located.

[0012] In a preferred embodiment, obtaining the compensated time delay of the target to be located includes: By performing deviation characteristic analysis on the dynamic deviation residual and the initial time delay estimate, the deviation compensation amount between the dynamic deviation residual and the initial time delay estimate is obtained; By establishing the compensation relationship for the deviation compensation amount, the compensation mapping relationship of the initial time delay estimate is obtained; Based on the compensation mapping relationship, the initial time delay estimate is subjected to time delay offset removal to obtain the compensated time delay of the target to be located.

[0013] In a preferred embodiment, obtaining the spatial location of the target to be located includes: By performing spatial domain mapping on the adaptive time delay correction, a candidate set of spatial locations for the target to be located is obtained; A consistency check is performed on the spatial location candidate set to obtain a filtered location candidate set for the target to be located. Spatial intersection analysis is performed on the filtered candidate location set to obtain the spatial intersection geometric relationship of the filtered candidate location set; Based on spatial intersection geometry, the selected candidate locations are fused to obtain the spatial location of the target to be located.

[0014] To address the above problems, the present invention also provides an adaptive underwater acoustic positioning correction system based on swarm intelligence, the system comprising: The high-dimensional phase space reconstruction module is used to reconstruct the original signals acquired by the underwater acoustic reference nodes during mutual detection in a high-dimensional phase space, and extract the arrival time feature vector of the original signals from the reconstructed phase domain trajectory. The multi-objective particle swarm optimization module is used to perform multi-objective optimization on the local neighborhood geometry of the arrival time feature vector based on the multi-objective particle swarm optimization strategy, so as to obtain the optimal local linear embedding parameters of the arrival time feature vector. The global manifold stitching module is used to perform global manifold stitching on the arrival time feature vector based on the optimal local linear embedding parameters to obtain the time delay spatial manifold of the underwater acoustic reference node. The variational mode decomposition module is used to perform variational mode decomposition on the time-delay spatial manifold to obtain the dynamic deviation residuals of the underwater acoustic reference node; The time delay distortion correction module is used to perform time delay distortion correction based on the projection position of the transmitted signal of the target to be located on the time delay space manifold and to fuse the dynamic deviation residual, so as to obtain the adaptive time delay correction amount of the target to be located. The intersection positioning and analysis module is used to perform intersection positioning and analysis on the adaptive time delay correction to obtain the spatial position of the target to be located.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention first reconstructs the original signal acquired by the underwater acoustic reference node into a high-dimensional phase space. Then, it determines the appropriate embedding dimension and delay time through traversal optimization, maps the signal into phase points, and traces these phase points in time sequence to form a complete phase domain trajectory. After obtaining this trajectory, it performs quantitative analysis on its local geometric features such as curvature, slope, and point density. This allows for the complete and accurate extraction of time-of-arrival features from the signal, forming a time-of-arrival feature vector. This approach fully extracts the hidden temporal features in the high-dimensional signal, making the data used in subsequent optimization steps more reliable and fundamentally improving the problems of insufficient feature extraction and low quality of basic data in traditional methods.

[0016] 2. This invention employs a multi-objective particle swarm optimization strategy, focusing on two core dimensions: reconstruction accuracy and topology preservation. It iteratively optimizes the local geometric neighborhood of the arrival time feature vector through multiple rounds, and performs topology consistency checks to obtain the optimal local linear embedding parameters. Then, a time-delay spatial manifold is constructed through global manifold stitching, and the dynamic deviation residuals are separated using variational mode decomposition. Finally, the adaptive time-delay correction is spatially mapped, and after consistency checks, the result is obtained through intersection localization analysis. This entire process forms a complete optimization chain that can accommodate various needs of underwater acoustic localization, ensuring that the optimized parameters are adaptable to all time-delay correction scenarios. It also effectively eliminates the distortion effects caused by environmental interference and equipment errors, significantly improving the correction efficiency and spatial positioning accuracy of adaptive underwater acoustic localization. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating an embodiment of the adaptive underwater acoustic localization correction method based on swarm intelligence provided by the present invention. Figure 2 A functional block diagram of an adaptive underwater acoustic positioning correction system based on swarm intelligence provided in an embodiment of the present invention; The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0018] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0019] This application provides an adaptive underwater acoustic positioning correction method based on swarm intelligence. The execution subject of the adaptive underwater acoustic positioning correction method based on swarm intelligence includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the adaptive underwater acoustic positioning correction method based on swarm intelligence can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0020] Reference Figure 1The diagram shown is a flowchart illustrating an adaptive underwater acoustic localization correction method based on swarm intelligence provided in an embodiment of the present invention. In this embodiment, the adaptive underwater acoustic localization correction method based on swarm intelligence includes: S1. The original signals collected by the underwater acoustic reference nodes during mutual detection are reconstructed in high-dimensional phase space, and the arrival time feature vector of the original signals is extracted from the reconstructed phase domain trajectory. In this embodiment of the invention, the extraction of the arrival time feature vector of the original signal includes: The phase space of the original signals acquired by the underwater acoustic reference nodes during mutual detection is reconstructed to obtain the high-dimensional phase space of the original signals; The phase space of the original signal collected by the underwater acoustic reference nodes during mutual detection is reconstructed. Based on the time series data of the original signal, the one-dimensional time series of the original signal is expanded by rearranging and combining the sampling data at different time points in the original signal according to a fixed time interval. The single-dimensional signal data is transformed into a multi-dimensional spatial data structure. Finally, a high-dimensional phase space of the original signal is obtained by reconstructing the time series data of the original signal. This high-dimensional phase space is a multi-dimensional data space that can reflect the characteristic distribution of the original signal in different time dimensions.

[0021] Based on the embedding dimension and delay time determined by traversal optimization, the embedding phase point mapping is performed on the high-dimensional phase space to obtain the phase point set of the original signal; Based on the embedding dimension and delay time determined by traversal optimization, embedding phase point mapping is performed on the high-dimensional phase space. First, all data points in the high-dimensional phase space are traversed, and the appropriate embedding dimension and delay time are determined according to the temporal correlation characteristics and spatial distribution characteristics of the data points. Then, the determined embedding dimension is used as the spatial dimension standard, and the delay time is used as the data point selection interval. The corresponding signal data points are extracted from the high-dimensional phase space and spatially mapped. The data points in the high-dimensional phase space are mapped to the specified phase space coordinate system according to the set dimension and interval. Finally, the phase point set of the original signal composed of all the mapped signal data points is obtained. This phase point set is the collection of signal data points in the high-dimensional phase space that meet the embedding dimension and delay time requirements.

[0022] Neighborhood trajectory tracing is performed on the set of phase points to obtain the phase domain trajectory of the original signal; To perform neighborhood trajectory tracking on a set of phase points, first determine the neighborhood range of each data point in the set of phase points. This neighborhood range is a fixed range centered on a single data point and including surrounding related data points. Then, according to the time sequence of the original signal, track the spatial movement trajectory of the related data points in the neighborhood of each data point in turn. The neighborhood trajectories of each data point are continuously spliced ​​together according to the time sequence to finally obtain the phase domain trajectory of the original signal, which can reflect the spatial trajectory of the original signal characteristics changing over time. This phase domain trajectory is a continuous spatial trajectory line of the data points in the set of phase points changing over time.

[0023] Local geometric feature analysis is performed on the phase domain trajectory, and the arrival time feature vector of the original signal is extracted from the analysis results; Local geometric feature analysis is performed on the phase domain trajectory, and the arrival time feature vector of the original signal is extracted from the analysis results. First, the phase domain trajectory is divided into multiple local trajectory segments according to fixed time segments. Then, the curvature, slope, direction, and data point density of each local trajectory segment are analyzed in sequence. Curvature is the degree of bending of the local trajectory segment, slope is the degree of tilt of the local trajectory segment, direction is the spatial extension direction of the local trajectory segment, and data point density is the number of signal data points per unit space within the local trajectory segment. Then, the analysis results of curvature, slope, direction, and data point density of each local trajectory segment are quantized. The quantized feature values ​​are arranged and combined in a set order. Finally, the arrival time feature vector of the original signal composed of quantized local geometric feature values ​​is extracted. This arrival time feature vector is a sequence of feature values ​​that can characterize the arrival time-related features of the original signal.

[0024] S2. Based on the multi-objective particle swarm optimization strategy, the local neighborhood geometry of the arrival time feature vector is optimized in multiple objectives to obtain the optimal local linear embedding parameters of the arrival time feature vector. In this embodiment of the invention, obtaining the optimal local linear embedding parameters of the arrival time feature vector includes: The arrival time feature vector is partitioned into its neighborhood to obtain the local geometric neighborhood of the arrival time feature vector; The arrival time feature vector is divided into regions according to the degree of correlation of spatial geometric distribution. Taking each feature point in the arrival time feature vector as the core, a fixed spatial neighborhood range is defined. The core point and all related feature points within the neighborhood range are grouped into an independent geometric region, thus completing the global partitioning of the arrival time feature vector. The resulting independent geometric region is the local geometric neighborhood of the arrival time feature vector. Each local geometric neighborhood contains the core feature point and a set of spatially related feature points around it.

[0025] Based on a multi-objective particle swarm optimization strategy, the quality of neighborhood reconstruction is evaluated for local geometric neighborhoods, and the fitness evaluation index of local geometric neighborhoods is obtained. Based on the core logic of multi-objective particle swarm optimization, this method uses the reconstruction accuracy and topology preservation of local geometric neighborhoods as the core evaluation dimensions. Reconstruction is performed on each local geometric neighborhood, and the reconstructed neighborhood is compared with the original neighborhood in terms of spatial features. The overlap between the two neighborhoods in terms of feature point distribution, geometric shape, and neighborhood relationships is calculated. Simultaneously, the consistency between the topological structure of the reconstructed neighborhood and the original structure is verified. The quantitative results of overlap and topological consistency are integrated to form a set of quantitative indicators that comprehensively reflect the reconstruction quality of local geometric neighborhoods. This set of quantitative indicators is the fitness evaluation index for local geometric neighborhoods. The value of the fitness evaluation index directly corresponds to the quality of local geometric neighborhood reconstruction; the higher the value, the better the reconstruction quality.

[0026] Based on the fitness evaluation index, the local geometric neighborhood is iteratively optimized and selected to obtain high-quality embedding candidates for the local geometric neighborhood; The fitness evaluation metrics of all local geometric neighborhoods are sorted from high to low. The top-ranked local geometric neighborhoods are selected as the initial selection objects. Using the preset threshold of the fitness evaluation metrics as the screening criterion, the initial selection objects are verified one by one. Local geometric neighborhoods that do not reach the preset threshold are eliminated. For local geometric neighborhoods that reach the preset threshold, multiple rounds of optimization iterations are performed. In each round of iteration, the reconstruction parameters of the local geometric neighborhood are fine-tuned and the fitness evaluation metrics are recalculated. Local geometric neighborhoods whose fitness evaluation metrics improve in each round of iteration are retained, while those whose metrics decrease are eliminated. After multiple rounds of iterations, all the local geometric neighborhoods that are finally retained are the high-quality embedding candidates. The high-quality embedding candidates are all local geometric neighborhoods whose reconstruction quality meets the preset criteria and has been optimized in multiple rounds.

[0027] Topological consistency verification is performed on high-quality embedding candidates to obtain stable embedding results in local geometric neighborhoods; The topological structure correlation of high-quality embedding candidates is verified globally. The spatial geometry, neighborhood correlation of feature points, and connection boundaries between regions are examined one by one. The consistency between the topological structure of each high-quality embedding candidate and the overall topological structure of the arrival time feature vector is verified. At the same time, the connection and compatibility of the topological structures between the high-quality embedding candidates are verified. High-quality embedding candidates whose topological structure is inconsistent with the whole or whose connection between regions is conflicting are eliminated. High-quality embedding candidates whose topological structure fully meets the overall requirements and whose connection between regions is smooth are retained. The embedding results corresponding to these retained high-quality embedding candidates are the stable embedding results of the local geometric neighborhood. The stable embedding results are the set of local embedding parameters with topological consistency and no structural conflicts.

[0028] Based on the stable embedding results, a global optimal matching is performed to obtain the optimal local linear embedding parameters for the arrival time feature vector; Stable embedding results from all local geometric neighborhoods are included in the global matching range. Using the overall spatial geometric distribution of the arrival time feature vector as the matching benchmark, the fit degree between each stable embedding result and the overall benchmark is calculated. At the same time, the co-fit degree between each stable embedding result is calculated. The fit degree and co-fit degree are weighted and integrated. The stable embedding result with the highest integrated value is selected as the basis. The parameters of the stable embedding results in the global range are fused and optimized to make the adjusted embedding parameters adaptable to all local geometric neighborhoods of the arrival time feature vector and maintain the consistency of geometric features and topology in the global range. The final set of optimized embedding parameters that can adapt to the global range is the optimal local linear embedding parameters for the arrival time feature vector.

[0029] S3. Based on the optimal local linear embedding parameters, the arrival time feature vectors are globally stitched together to obtain the time delay spatial manifold of the underwater acoustic reference node; In this embodiment of the invention, obtaining the time-delay spatial manifold of the underwater acoustic reference node includes: Based on the optimal local linear embedding parameters, local linear embedding is performed on the arrival time feature vector to obtain the local low-dimensional embedding coordinates of the arrival time feature vector; Based on the optimal local linear embedding parameters, feature dimensions with linear correlation characteristics are selected from the arrival time feature vectors. The high-dimensional arrival time feature vectors are linearly projected onto the low-dimensional space. During the projection process, the local geometric structure features and data correlation relationships of the arrival time feature vectors are preserved. The set of low-dimensional space coordinates formed after the projection transformation is the local low-dimensional embedding coordinates of the arrival time feature vectors. These coordinates are the low-dimensional data coordinates obtained after dimensionality reduction of the high-dimensional arrival time feature vectors. Each feature vector corresponds to a unique local low-dimensional embedding coordinate.

[0030] Based on the optimal local linear embedding parameters, the local low-dimensional embedding coordinates are globally concatenated to obtain a global permutation matrix of arrival time feature vectors; Based on the topological association rules of local neighborhoods in the optimal local linear embedding parameters, the local low-dimensional embedding coordinates of all arrival time feature vectors are arranged in an orderly manner according to the detection time order of the underwater acoustic reference nodes and the spatial position association of the nodes. The arranged local low-dimensional embedding coordinates are integrated in the form of a matrix, where the rows and columns of the matrix correspond to the dimensions of the local low-dimensional embedding coordinates and the sequence of feature vectors, respectively. The two-dimensional matrix structure formed after integration is the global arrangement matrix of arrival time feature vectors. This matrix is ​​a structured data carrier after the orderly integration of all local low-dimensional embedding coordinates, which fully preserves the spatiotemporal and topological associations between the coordinates.

[0031] Perform eigenspectral decomposition on the global permutation matrix to obtain the low-dimensional eigencoordinates of the global permutation matrix; The eigenvalues ​​and eigenvectors of the global permutation matrix are solved. The eigenvectors corresponding to the eigenvalues ​​that accumulate to a preset high percentage after sorting the eigenvalues ​​from largest to smallest are selected. The selected eigenvectors are used as the basis of the low-dimensional space. All data in the global permutation matrix are projected onto this basis. The coordinate set obtained after projection is the low-dimensional intrinsic coordinate of the global permutation matrix. This coordinate is the core low-dimensional coordinate obtained after feature extraction and dimensionality reduction of the global permutation matrix, which retains the main data features and structural information of the global permutation matrix.

[0032] Based on the low-dimensional intrinsic coordinates, the time delay observation data of the underwater acoustic reference node are calibrated using manifold coordinates to obtain the time delay spatial manifold of the underwater acoustic reference node; Using low-dimensional intrinsic coordinates as a spatial reference, all time-delay observation data collected by underwater acoustic reference nodes during mutual detection are matched one-to-one with the data acquisition time, the corresponding reference node identifier, and the low-dimensional intrinsic coordinates. A unique spatial position is determined for each time-delay observation data in the space formed by the low-dimensional intrinsic coordinates. All time-delay observation data with completed position calibration are integrated in this space, and the resulting continuous spatial structure that reflects the spatial distribution and variation law of time-delay observation data is the time-delay spatial manifold of the underwater acoustic reference node. This manifold is the concrete spatial distribution form of time-delay observation data in the low-dimensional intrinsic coordinate space, which fully presents the overall characteristics and local correlations of the time-delay data of the underwater acoustic reference node.

[0033] S4. Perform variational mode decomposition on the time-delayed spatial manifold to obtain the dynamic deviation residuals of the underwater acoustic reference node; In this embodiment of the invention, obtaining the dynamic deviation residual of the underwater acoustic reference node includes: Perform decomposition parameter initialization on the time-delay space manifold to obtain the number of modal decompositions and the penalty factor of the time-delay space manifold; For the time-delay spatial manifold, the decomposition parameter initialization operation is carried out. First, all time-delay observation data in the time-delay spatial manifold are extracted. This data is the time-delay related data collected by the underwater acoustic reference nodes during mutual detection and calibrated by the manifold coordinates. Then, based on the overall data volume and the complexity of the data distribution of the time-delay spatial manifold, the number of suitable modal decompositions is determined. This number is the number of intrinsic modal components obtained after decomposing the time-delay spatial manifold. At the same time, based on the data fluctuation amplitude and data noise level of the time-delay spatial manifold, a penalty factor is determined. This factor is used to constrain the process of intrinsic mode iterative separation. Finally, the number of modal decompositions and the penalty factor that can adapt to the decomposition requirements of the time-delay spatial manifold are obtained.

[0034] Based on the number of modal decompositions and the penalty factor, the intrinsic modal components of the time-delay spatial manifold are obtained by iterative separation of the intrinsic modes. Based on the determined number of modal decompositions and penalty factors, intrinsic modal iterative separation is performed on the time-delay spatial manifold. First, the overall data of the time-delay spatial manifold is initially divided according to the number of modal decompositions. Then, with the penalty factor as a constraint, the data groups after the initial division are repeatedly separated and verified. In each iteration, it is verified whether the separated components meet the characteristic requirements of intrinsic modes. If they do not meet the requirements, the data division method is adjusted and separation continues until the set number of iterations is completed and all separated components meet the characteristic requirements of intrinsic modes. Finally, the intrinsic modal components formed by the iterative separation of the time-delay spatial manifold are obtained. These components are modal data units with independent characteristics in the time-delay spatial manifold.

[0035] Residual component identification is performed on the intrinsic modal components to obtain the trend term components of the time-delay spatial manifold, including: The intrinsic modal components are evaluated for trend characteristics to obtain trend evaluation indices for the intrinsic modal components. To assess the trend characteristics of intrinsic modal components, firstly, all data points for each intrinsic modal component are extracted and the time series relationships of each data point are analyzed. Then, the overall amplitude and stability of the direction of change of each intrinsic modal component over time are calculated. The amplitude of change is determined by the range of numerical changes of data points per unit time, and the stability of the direction of change is determined by whether the increasing or decreasing trend of data point values ​​is consistent within a continuous time period. Subsequently, based on the magnitude of the amplitude of change and the stability of the direction of change, a corresponding quantitative value is assigned to each intrinsic modal component. This quantitative value is the trend assessment index that can characterize the trend characteristics of the intrinsic modal component.

[0036] The trend assessment indicators are subjected to trend mode screening to obtain the trend mode components of the trend assessment indicators; The obtained trend evaluation indicators are used to screen trend modes. First, a screening threshold for the trend evaluation indicators is set. This threshold is determined based on the overall trend characteristics of the time-delay spatial manifold. Then, the trend evaluation indicator corresponding to each intrinsic mode component is compared with the screening threshold. Intrinsic mode components with evaluation indicator values ​​higher than the screening threshold are retained, while intrinsic mode components with evaluation indicator values ​​lower than the screening threshold are removed. Finally, the trend mode components with obvious trend characteristics are obtained after screening.

[0037] By integrating and reconstructing the trend modal components, the trend term components of the time-delay spatial manifold are obtained; The trend modal components are integrated and reconstructed. First, all trend modal components are aligned according to the time series so that the values ​​of each trend modal component at the same time node can match each other. Then, the values ​​of the corresponding time nodes of the aligned trend modal components are superimposed and calculated. The superimposed value is used as the value of the reconstructed data point. Then, all the reconstructed data points are arranged continuously according to the time series. Finally, the trend component that can characterize the overall trend characteristics of the time-delayed spatial manifold is obtained. This component is the core data part of the time-delayed spatial manifold that shows regular changes over time.

[0038] Trend term deviation analysis is performed on the trend term components to obtain the dynamic deviation residuals of the underwater acoustic reference node; Trend term deviation analysis is performed on the trend term components. First, all data points of the trend term components are extracted and an ideal trend curve of the trend term components is fitted. This curve is an unbiased standard trend curve obtained according to the overall change law of the trend term components. Then, the difference between the actual value of each data point in the trend term components and the standard value at the corresponding position on the ideal trend curve is calculated. All differences are integrated according to the time series, and finally, the dynamic deviation residual that can characterize the trend deviation in the time delay spatial manifold is obtained. This residual is the dynamic time delay deviation data generated by the underwater acoustic reference node during the detection process.

[0039] S5. Based on the projection position of the transmitted signal of the target to be located on the time delay space manifold, and by fusing the dynamic deviation residual to perform time delay distortion correction, the adaptive time delay correction amount of the target to be located is obtained. In this embodiment of the invention, obtaining the adaptive time delay correction amount of the target to be located includes: Perform manifold projection positioning on the transmitted signal of the target to be located to obtain the geometric projection position of the transmitted signal on the time delay space manifold; The specific operation of manifold projection localization for the transmitted signal of the target to be located involves first acquiring the full time-domain data of the transmitted signal of the target to be located. This data is a set of sampled data of the transmitted signal of the target to be located received by the underwater acoustic reference node at different time points. Then, the time-domain data is matched with the coordinate dimension according to the coordinate calibration rules of the time-delay space manifold, and the time-domain data of the transmitted signal is transformed into vector data consistent with the coordinate dimension of the time-delay space manifold. Subsequently, the vector data is projected into the coordinate system of the time-delay space manifold. By calculating the spatial distance between the vector data and each coordinate point on the time-delay space manifold, the coordinate point with the smallest distance is selected as the geometric projection position of the transmitted signal on the time-delay space manifold. This geometric projection position is the unique spatial coordinate point of the transmitted signal of the target to be located in the coordinate system of the time-delay space manifold.

[0040] Based on the geometric projection position, the initial time delay of the transmitted signal is extracted to obtain the initial time delay estimate of the target to be located; The specific operation for extracting the initial time delay of the transmitted signal based on the geometric projection position is as follows: First, retrieve the time delay reference data corresponding to the geometric projection position in the time delay space manifold. This data is the signal propagation time delay data corresponding to each coordinate point pre-calibrated when the underwater acoustic reference node constructs the time delay space manifold. Then, combine the signal propagation path information corresponding to the underwater acoustic reference node and the geometric projection position, extract the time delay value corresponding to the geometric projection position under the propagation path, and use this value as the initial time delay estimate of the target to be located. This initial time delay estimate is the time delay value of the transmitted signal of the target to be located reaching the underwater acoustic reference node without deviation correction.

[0041] Based on the dynamic bias residual, the initial time delay estimate is corrected by bias compensation to obtain the compensated time delay of the target to be located, including: By performing deviation characteristic analysis on the dynamic deviation residual and the initial time delay estimate, the deviation compensation amount between the dynamic deviation residual and the initial time delay estimate is obtained; The specific operation of deviation feature analysis between dynamic deviation residuals and initial time delay estimates involves first acquiring the full data of dynamic deviation residuals, which is a set of deviation values ​​obtained after variational mode decomposition of the time delay spatial manifold. Then, the dynamic deviation residual data and the initial time delay estimate are aligned according to the time series of signal propagation. The numerical differences between the two at each stage of signal propagation are analyzed point by point, and the variation law and characteristic performance of the difference values ​​are extracted. Based on the characteristic performance, a specific value that can compensate for the deviation of the initial time delay estimate is calculated. This value is used as the deviation compensation amount between the dynamic deviation residuals and the initial time delay estimate. This deviation compensation amount is the specific value used to correct the deviation of the initial time delay estimate.

[0042] By establishing the compensation relationship for the deviation compensation amount, the compensation mapping relationship of the initial time delay estimate is obtained; The specific operation for establishing the compensation relationship for the deviation compensation amount is to first conduct a quantitative analysis of the numerical correlation between the deviation compensation amount and the initial time delay estimate, clarify the corresponding change law of the deviation compensation amount as the numerical value of the initial time delay estimate changes, and then establish a one-to-one correspondence between the initial time delay estimate and the deviation compensation amount based on this change law. This correspondence is used as the compensation mapping relationship of the initial time delay estimate. This compensation mapping relationship is a numerical correspondence rule that can clarify which deviation compensation amount should be matched with the initial time delay estimate.

[0043] Based on the compensation mapping relationship, the initial time delay estimate is subjected to time delay offset removal to obtain the compensated time delay of the target to be located. The specific operation of removing time delay offsets from the initial time delay estimate based on the compensation mapping relationship is as follows: First, match the corresponding deviation compensation amount in the compensation mapping relationship according to the value of the initial time delay estimate. Then, perform numerical fusion processing on the initial time delay estimate and the matched deviation compensation amount to remove the time delay offset value caused by environmental interference, equipment error and other factors in the initial time delay estimate. The time delay value after offset removal is obtained and used as the compensated time delay value of the target to be located. The compensated time delay value is the signal transmission delay value of the target to be located after deviation compensation correction.

[0044] The time delay distortion is corrected by applying time delay correction to the compensated time delay amount to obtain the adaptive time delay correction amount for the target to be located; The specific operation of correcting the time delay distortion of the compensated time delay is as follows: First, analyze the influence of factors such as temperature, salinity, and water flow in the underwater acoustic propagation environment on the distortion of signal propagation time delay, extract the time delay distortion characteristics corresponding to each influencing factor, and then correct the distortion value of the compensated time delay point by point according to the characteristics. Remove the distortion value caused by environmental factors in the compensated time delay to obtain the corrected time delay value. This value is used as the adaptive time delay correction value of the target to be located. This adaptive time delay correction value is the final time delay value of the transmitted signal of the target to be located after completing the deviation compensation and distortion correction.

[0045] S6. Perform intersection positioning analysis on the adaptive time delay correction to obtain the spatial position of the target to be located; In this embodiment of the invention, obtaining the spatial location of the target to be located includes: By performing spatial domain mapping on the adaptive time delay correction, a candidate set of spatial locations for the target to be located is obtained; Based on the adaptive time delay correction in the underwater acoustic positioning scenario, and combined with the known spatial coordinate information of the underwater acoustic reference node, a mapping relationship model between time delay and spatial position is established. Each adaptive time delay correction is substituted into this mapping relationship model, and the corresponding spatial coordinate data is obtained by solving the spatial coordinates. All the solved spatial coordinate data are integrated and collected to form a spatial position candidate set of the target to be located. This spatial position candidate set is a dataset composed of multiple spatial coordinates of the target to be located calculated based on different adaptive time delay corrections.

[0046] A consistency check is performed on the spatial location candidate set to obtain a filtered location candidate set for the target to be located. Each spatial coordinate data in the candidate spatial location set of the target to be located is compared with the preset reasonable range of spatial location in the underwater acoustic positioning system. At the same time, the reasonableness of the spatial distance between each spatial coordinate data and the spatial coordinate of the reference node is calculated. Spatial coordinate data that exceeds the preset reasonable range of spatial location are eliminated, as are spatial coordinate data whose spatial distance does not conform to the physical laws of underwater acoustic propagation. The spatial coordinate data retained after the above double screening are re-integrated to form the filtered candidate location set of the target to be located. This filtered candidate location set is the valid spatial coordinate dataset retained after the spatial location candidate set has been verified for reasonableness.

[0047] Spatial intersection analysis is performed on the filtered candidate location set to obtain the spatial intersection geometric relationship of the filtered candidate location set; For all valid spatial coordinate data in the filtered candidate location set, a spatial sphere is constructed with each spatial coordinate as the origin and the positioning accuracy of the underwater acoustic positioning system as the radius. The spatial intersection region between any two spatial spheres is calculated, and then all spatial intersection regions are superimposed to determine the intersection, tangency, or separation states between the spatial spheres corresponding to each valid spatial coordinate. At the same time, the number and spatial range of spatial coordinates contained in each intersection region are counted to determine the spatial intersection geometric relationship of the filtered candidate location set. This spatial intersection geometric relationship is a comprehensive representation of the spatial position association characteristics and intersection region characteristics of the spatial spheres corresponding to the valid spatial coordinates in the filtered candidate location set.

[0048] Based on spatial intersection geometry, the selected candidate locations are fused to obtain the spatial location of the target to be located; Based on the spatial intersection geometric relationships of the selected candidate locations, the common intersection region of all spatial spheres is extracted, and the geometric center coordinates of this common intersection region are calculated. Using the geometric center coordinates as the basis, a weighted average is calculated for all valid spatial coordinate data within the common intersection region. Spatial coordinates with a larger proportion of the intersection region in the spatial intersection geometric relationship are assigned higher weights, and spatial coordinates with a smaller proportion of the intersection region are assigned lower weights. The final spatial coordinate data obtained after weighted average calculation is determined as the spatial location of the target to be located. This spatial location is a three-dimensional spatial coordinate that can accurately represent the actual location of the target.

[0049] like Figure 2 The diagram shown is a functional block diagram of an adaptive underwater acoustic positioning correction system based on swarm intelligence provided in an embodiment of the present invention.

[0050] The adaptive underwater acoustic positioning correction system 100 based on swarm intelligence described in this invention can be installed in an electronic device. Depending on the functions implemented, the adaptive underwater acoustic positioning correction system 100 based on swarm intelligence may include a high-dimensional phase space reconstruction module 101, a multi-objective particle swarm optimization module 102, a manifold global stitching module 103, a variational mode decomposition module 104, a time delay distortion correction module 105, and an intersection positioning analysis module 106. The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and are stored in the memory of the electronic device.

[0051] In this embodiment, the functions of each module / unit are as follows: The high-dimensional phase space reconstruction module 101 is used to reconstruct the original signals acquired by the underwater acoustic reference nodes during mutual detection in a high-dimensional phase space, and to extract the arrival time feature vector of the original signals from the reconstructed phase domain trajectory. The multi-objective particle swarm optimization module 102 is used to perform multi-objective optimization on the local neighborhood geometry of the arrival time feature vector based on the multi-objective particle swarm optimization strategy, so as to obtain the optimal local linear embedding parameters of the arrival time feature vector. The manifold global stitching module 103 is used to perform manifold global stitching on the arrival time feature vector based on the optimal local linear embedding parameters to obtain the time delay spatial manifold of the underwater acoustic reference node. The variational mode decomposition module 104 is used to perform variational mode decomposition on the time-delay spatial manifold to obtain the dynamic deviation residuals of the underwater acoustic reference node. The time delay distortion correction module 105 is used to perform time delay distortion correction based on the projection position of the transmitted signal of the target to be located on the time delay space manifold and to fuse the dynamic deviation residual, so as to obtain the adaptive time delay correction amount of the target to be located. The intersection positioning and analysis module 106 is used to perform intersection positioning and analysis on the adaptive time delay correction to obtain the spatial position of the target to be located.

[0052] In the several embodiments provided by the present invention, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the division of modules is merely a logical functional division, and there may be other division methods in actual implementation.

[0053] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules may be selected to achieve the purpose of this embodiment according to actual needs.

[0054] In addition, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0055] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0056] This application embodiment can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0057] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An adaptive underwater acoustic localization correction method based on swarm intelligence, characterized in that, The method includes: S1. The original signals collected by the underwater acoustic reference nodes during mutual detection are reconstructed in high-dimensional phase space, and the arrival time feature vector of the original signals is extracted from the reconstructed phase domain trajectory. S2. Based on the multi-objective particle swarm optimization strategy, the local neighborhood geometry of the arrival time feature vector is optimized in multiple objectives to obtain the optimal local linear embedding parameters of the arrival time feature vector. S3. Based on the optimal local linear embedding parameters, the arrival time feature vectors are globally stitched together to obtain the time delay spatial manifold of the underwater acoustic reference node; S4. Perform variational mode decomposition on the time-delayed spatial manifold to obtain the dynamic deviation residuals of the underwater acoustic reference node; S5. Based on the projection position of the transmitted signal of the target to be located on the time delay space manifold, and by fusing the dynamic deviation residual to perform time delay distortion correction, the adaptive time delay correction amount of the target to be located is obtained. S6. Perform intersection positioning analysis on the adaptive time delay correction to obtain the spatial position of the target to be located.

2. The adaptive underwater acoustic localization correction method based on swarm intelligence as described in claim 1, characterized in that, The extraction of the arrival time feature vector of the original signal includes: The phase space of the original signals acquired by the underwater acoustic reference nodes during mutual detection is reconstructed to obtain the high-dimensional phase space of the original signals; Based on the embedding dimension and delay time determined by traversal optimization, embedding phase point mapping is performed on the high-dimensional phase space to obtain the phase point set of the original signal; Neighborhood trajectory tracing is performed on the set of phase points to obtain the phase domain trajectory of the original signal; Local geometric feature analysis is performed on the phase domain trajectory, and the arrival time feature vector of the original signal is extracted from the analysis results.

3. The adaptive underwater acoustic localization correction method based on swarm intelligence as described in claim 1, characterized in that, The optimal local linear embedding parameters for obtaining the arrival time feature vector include: The arrival time feature vector is partitioned into its neighborhood to obtain the local geometric neighborhood of the arrival time feature vector; Based on a multi-objective particle swarm optimization strategy, the quality of neighborhood reconstruction is evaluated for local geometric neighborhoods, and the fitness evaluation index of local geometric neighborhoods is obtained. Based on the fitness evaluation index, the local geometric neighborhood is iteratively optimized and selected to obtain high-quality embedding candidates for the local geometric neighborhood; Topological consistency verification is performed on high-quality embedding candidates to obtain stable embedding results in local geometric neighborhoods; Based on the stable embedding results, a global optimal matching is performed to obtain the optimal local linear embedding parameters for the arrival time feature vector.

4. The adaptive underwater acoustic localization correction method based on swarm intelligence as described in claim 1, characterized in that, The obtained time-delay spatial manifold of the underwater acoustic reference node includes: Based on the optimal local linear embedding parameters, local linear embedding is performed on the arrival time feature vector to obtain the local low-dimensional embedding coordinates of the arrival time feature vector; Based on the optimal local linear embedding parameters, the local low-dimensional embedding coordinates are globally concatenated to obtain a global permutation matrix of arrival time feature vectors; Perform eigenspectral decomposition on the global permutation matrix to obtain the low-dimensional eigencoordinates of the global permutation matrix; Based on the low-dimensional intrinsic coordinates, the time delay observation data of the underwater acoustic reference node are calibrated using manifold coordinates to obtain the time delay spatial manifold of the underwater acoustic reference node.

5. The adaptive underwater acoustic localization correction method based on swarm intelligence as described in claim 1, characterized in that, The obtained dynamic deviation residual of the underwater acoustic reference node includes: Perform decomposition parameter initialization on the time-delay space manifold to obtain the number of modal decompositions and the penalty factor of the time-delay space manifold; Based on the number of modal decompositions and the penalty factor, the intrinsic modal components of the time-delay spatial manifold are obtained by iterative separation of the intrinsic modes. Residual component identification is performed on the intrinsic modal components to obtain the trend term components of the time-delay spatial manifold; Trend term deviation analysis is performed on the trend term components to obtain the dynamic deviation residuals of the underwater acoustic reference node.

6. The adaptive underwater acoustic localization correction method based on swarm intelligence as described in claim 5, characterized in that, The trend term components of the obtained time-delayed spatial manifold include: The intrinsic modal components are evaluated for trend characteristics to obtain trend evaluation indices for the intrinsic modal components. The trend assessment indicators are subjected to trend mode screening to obtain the trend mode components of the trend assessment indicators; By integrating and reconstructing the trend modal components, the trend term components of the time-delayed spatial manifold are obtained.

7. The adaptive underwater acoustic localization correction method based on swarm intelligence as described in claim 1, characterized in that, The process of obtaining the adaptive time delay correction amount for the target to be located includes: Perform manifold projection positioning on the transmitted signal of the target to be located to obtain the geometric projection position of the transmitted signal on the time delay space manifold; Based on the geometric projection position, the initial time delay of the transmitted signal is extracted to obtain the initial time delay estimate of the target to be located; Based on the dynamic deviation residual, the initial time delay estimate is corrected by deviation compensation to obtain the compensated time delay of the target to be located. The time delay is corrected for distortion after compensation to obtain the adaptive time delay correction for the target to be located.

8. The adaptive underwater acoustic localization correction method based on swarm intelligence as described in claim 7, characterized in that, The compensated time delay of the target to be located includes: By performing deviation characteristic analysis on the dynamic deviation residual and the initial time delay estimate, the deviation compensation amount between the dynamic deviation residual and the initial time delay estimate is obtained; By establishing the compensation relationship for the deviation compensation amount, the compensation mapping relationship of the initial time delay estimate is obtained; Based on the compensation mapping relationship, the initial time delay estimate is subjected to time delay offset removal to obtain the compensated time delay of the target to be located.

9. The adaptive underwater acoustic localization correction method based on swarm intelligence as described in claim 1, characterized in that, Obtaining the spatial location of the target to be located includes: By performing spatial domain mapping on the adaptive time delay correction, a candidate set of spatial locations for the target to be located is obtained; A consistency check is performed on the spatial location candidate set to obtain a filtered location candidate set for the target to be located. Spatial intersection analysis is performed on the filtered candidate location set to obtain the spatial intersection geometric relationship of the filtered candidate location set; Based on spatial intersection geometry, the selected candidate locations are fused to obtain the spatial location of the target to be located.

10. An adaptive underwater acoustic positioning correction system based on swarm intelligence, characterized in that, The system for implementing the adaptive underwater acoustic localization correction method based on swarm intelligence as described in claim 1 includes: The high-dimensional phase space reconstruction module is used to reconstruct the original signals acquired by the underwater acoustic reference nodes during mutual detection in a high-dimensional phase space, and extract the arrival time feature vector of the original signals from the reconstructed phase domain trajectory. The multi-objective particle swarm optimization module is used to perform multi-objective optimization on the local neighborhood geometry of the arrival time feature vector based on the multi-objective particle swarm optimization strategy, so as to obtain the optimal local linear embedding parameters of the arrival time feature vector. The global manifold stitching module is used to perform global manifold stitching on the arrival time feature vector based on the optimal local linear embedding parameters to obtain the time delay spatial manifold of the underwater acoustic reference node. The variational mode decomposition module is used to perform variational mode decomposition on the time-delay spatial manifold to obtain the dynamic deviation residuals of the underwater acoustic reference node; The time delay distortion correction module is used to perform time delay distortion correction based on the projection position of the transmitted signal of the target to be located on the time delay space manifold and to fuse the dynamic deviation residual, so as to obtain the adaptive time delay correction amount of the target to be located. The intersection positioning and analysis module is used to perform intersection positioning and analysis on the adaptive time delay correction to obtain the spatial position of the target to be located.

Images