A Cross-Location Method Based on Energy Fusion and Density Clustering

Abstract

The present application relates to a cross - location method based on energy fusion and density clustering, comprising the following steps: Step 1, obtaining the array received signals and generating a one - dimensional azimuth energy estimation spectrum; Step 2, dividing the region of interest into grids, and inversely mapping the one - dimensional azimuth energy estimation spectra of each array to the two - dimensional grid region in combination with the propagation loss; Step 3, fusing the two - dimensional energy spectra of each array obtained in Step 2 to obtain a multi - array fusion energy spectrum on the two - dimensional grid region. By inversely mapping the one - dimensional azimuth energy estimation spectrum generated from the multi - array received signals to the two - dimensional grid region in combination with the signal propagation loss model to form the two - dimensional energy spectra of each array, and adopting a fusion strategy to generate a two - dimensional fusion energy spectrum, the present invention significantly enhances the target signal features and suppresses noise interference, effectively reduces the false target points in cross - location, improves the positioning accuracy and reliability, and at the same time realizes automated processing and reduces the dependence on manual intervention.

Description

Technical Field

[0001] This application relates to the technical field of underwater acoustic signal positioning, and particularly to a cross-positioning method based on energy fusion and density clustering. Background Art

[0002] Underwater acoustic signal positioning is an important research direction in the field of underwater acoustic signal processing. Its core task is to accurately estimate the target position through the reception and analysis of the target radiated acoustic signal. As a key device for underwater target detection, the passive sonar system realizes target positioning by receiving the acoustic signal radiated by the target and extracting its characteristic parameters. Among them, the direction parameter (azimuth angle), as one of the most reliable parameters of the target signal source, becomes an important basis for positioning. The direction-finding cross technology based on the direction parameter is widely used in the field of underwater target detection and positioning due to its simple principle and easy implementation.

[0003] However, in practical applications, the complexity and uncertainty of the underwater environment bring many challenges to acoustic signal positioning. First, the signal is affected by factors such as multipath effect, noise interference, and water body absorption during the propagation process, resulting in a decrease in the signal-to-noise ratio of the received signal and a decrease in the accuracy of azimuth estimation. Second, when performing cross-positioning only through azimuth estimation, due to the existence of direction-finding errors, a large number of false target points often appear in the positioning results. Especially in a multi-target scenario, the intersection of direction-finding data of different signal sources will further exacerbate the generation of false positioning points. In addition, when the number of arrays exceeds two, the increase in the number of direction-finding lines will not only increase the computational complexity but also affect the positioning accuracy due to the accumulation of direction-finding errors. The existence of false positioning points will seriously interfere with the extraction of real targets and reduce the reliability and practicality of the positioning results.

[0004] Traditional cross-positioning methods usually rely on manual intervention or simple threshold segmentation techniques to extract the target position. This method is not only inefficient but also difficult to cope with the complex and changeable actual environment and cannot achieve automatic processing. Therefore, how to effectively suppress false positioning points and improve the accuracy and reliability of target positioning in a complex environment with high noise and multiple targets has become an urgent problem to be solved in the field of underwater acoustic signal positioning. Summary of the Invention

[0005] The purpose of the present invention is to provide a cross-positioning method based on energy fusion and density clustering in view of the deficiencies of the existing solutions. The present invention reduces the false positioning points in multi-array cross-positioning, improves the accuracy of target positioning, and reduces manual intervention, so as to automatically extract the target position more accurately.

[0006] A high-water-carrying humidifying wheel provided by this application adopts the following technical solutions:

[0007] A cross - location method based on energy fusion and density clustering, comprising the following steps:

[0008] Step 1: Obtain the array received signals and generate a one - dimensional azimuth energy estimation spectrum;

[0009] Step 2: Divide the region of interest into grids, and inversely map the one - dimensional azimuth energy estimation spectra of each array to the two - dimensional grid region in combination with the propagation loss;

[0010] Step 3: Fuse the two - dimensional energy spectra of each array obtained in Step 2 to obtain a multi - array fusion energy spectrum on the two - dimensional grid region;

[0011] Step 4: Extract potential target points on the fusion energy spectrum using an adaptive threshold technique according to the characteristics of the target points;

[0012] Step 5: Obtain the target location using an adaptive density clustering method according to the cross - location characteristics and the characteristics of the potential target points.

[0013] By adopting the above technical solution, through the fusion and clustering processing of multi - array data, it can effectively suppress noise and false target points, significantly improve the accuracy and reliability of target location. This method reduces the dependence on manual intervention, realizes the automation of the location process, and is applicable to complex and changeable underwater environments.

[0014] Optionally, the step of inversely mapping the one - dimensional azimuth energy estimation spectra of each array to the two - dimensional grid region in combination with the propagation loss in Step 2 is as follows:

[0015] Assume that the region of interest is , and the central point position of each array is ;

[0016] Calculate the angles and distances of each grid point relative to each array:

[0017]

[0018]

[0019] Calculate the propagation loss of the grid point to each array:

[0020]

[0021] Among them, is the propagation loss coefficient;

[0022] Calculate the two - dimensional energy spectrum of each array:

[0023]

[0024] Among them, is the The one-dimensional azimuth energy estimation spectrum of an array, is the grid point at the position relative to the th array's propagation loss coefficient.

[0025] By adopting the above technical solution, the region of interest is meshed and combined with the propagation loss model, and the one-dimensional azimuth energy estimation spectrum is inverted into a two-dimensional grid region, realizing the accurate mapping of the target energy in space, being able to better adapt to the signal attenuation characteristics in the underwater complex environment, and providing a reliable data basis for subsequent energy fusion and positioning.

[0026] Optionally, the fusion formula in step three is:

[0027]

[0028] where, is the two-dimensional multi-array fusion energy spectrum, is the th array's two-dimensional energy spectrum.

[0029] By adopting the above technical solution, the multi-array energy spectra are weighted and fused, enhancing the energy characteristics of the target signal, while suppressing noise and interference, making the fused energy spectrum able to highlight the local minimum characteristics of the target point, facilitating subsequent target extraction, and effectively reducing the false target points caused by the intersection of multiple bearing lines.

[0030] Optionally, the steps of extracting potential target points according to the characteristics of the target points on the fusion energy spectrum in step four are:

[0031] A1. Extract the ridge line energy data in the fusion energy spectrum along the maximum value angle in the one-dimensional azimuth estimation spectrum;

[0032] A2. Use polynomial fitting to remove the ridge line trend to obtain the detrended ridge line data ;

[0033] A3. Combine the Pauta criterion and set the threshold to , where , to obtain the potential target points that meet the conditions:

[0034] .

[0035] By adopting the above technical solutions, the adaptive threshold technology is used to extract potential target points, which can dynamically adjust the threshold according to the characteristics of the fusion energy spectrum to meet the target detection requirements in different environments. Moreover, through the extraction and detrending processing of the ridge energy data and setting the threshold in combination with the Pauta criterion, the accuracy of target extraction is significantly improved, and the noise interference is effectively suppressed.

[0036] Optionally, the steps of obtaining the target position by using the adaptive density clustering method in step five are as follows:

[0037] B1. Set the minimum neighborhood number to M - 1, where M is the number of arrays;

[0038] B2. According to the main lobe widths at different angles of each array and in combination with the distance from the potential target point to the array, set different neighborhood distances for different potential target points:

[0039]

[0040] where, is the distance from the potential target point ( to the center point of the array, is the angle formed by the potential target point and the array;

[0041] B3. Traverse all potential target points according to the density clustering method, and divide the potential target points with a distance less than the neighborhood distance into the same set. If the number of potential target points in the set is greater than the minimum neighborhood number, it is a positioning cluster; otherwise, it is an interference cluster;

[0042] B4. According to the clusters obtained by clustering, the position of the positioning point is the mean value of the potential target points in each cluster:

[0043] .

[0044] By adopting the above technical solutions, the potential target points are divided into positioning clusters and interference clusters, effectively distinguishing real targets from false targets, clustering according to the spatial distribution density of the potential target points, ensuring the reliability of the positioning results, especially suitable for multi-target scenarios, and avoiding target confusion.

[0045] Optionally, the signal received by the array is an active sonar signal, a passive sonar signal, or an underwater communication signal.

[0046] By adopting the above technical solutions, it is used to clarify the signal type, can better adapt to the particularity of the underwater environment, give full play to the propagation advantages of sound waves in water, and improve the accuracy and reliability of target detection and positioning.

[0047] Optionally, when the signal expands in the form of a cylindrical wave, the energy attenuation is inversely proportional to the distance, and at this time = 1;

[0048] When the signal expands in the form of a spherical wave, the energy attenuation is inversely proportional to the square of the distance. At this time = 2.

[0049] By adopting the above technical solution, it is used to adapt to different signal crosstalk modes and conform to the signal attenuation characteristics in the actual environment, and further improve the positioning accuracy.

[0050] Optionally, when in an environment with high underwater noise, increase the value to increase the threshold and reduce the false detection of noise points;

[0051] When in an environment with weak target signals, reduce the value to lower the threshold and avoid missing the detection of the target.

[0052] By adopting the above technical solution, it is used to flexibly adjust the threshold according to the actual application scenario to improve the accuracy and adaptability of target detection.

[0053] Optionally, in step five, the density clustering method is the DBSCAN or OPTICS clustering algorithm.

[0054] By adopting the above technical solution, the most suitable clustering method can be selected according to the actual needs, enhancing the applicability of this positioning method to facilitate the processing of target point sets with different scales and distributions.

[0055] In summary, this application includes at least one of the following beneficial technical effects:

[0056] The present invention combines the one-dimensional azimuth energy estimation spectrum generated by multi-array received signals with the signal propagation loss model and inversely maps it to a two-dimensional grid area to form the two-dimensional energy spectra of each array, and adopts a fusion strategy to generate a two-dimensional fusion energy spectrum, significantly enhancing the target signal characteristics and suppressing noise interference, effectively reducing the false target points in cross positioning, improving the positioning accuracy and reliability, and at the same time realizing automatic processing, reducing the dependence on manual intervention. This method can meet the core application scenarios of underwater acoustic positioning, adapt to the multi-target detection requirements in complex environments, and provide an efficient and reliable technical means for underwater target positioning. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 is a flowchart of the cross-positioning method based on energy fusion and density clustering of the present invention;

[0058] Figure 2 is the two-dimensional energy spectrum of the first array before fusion of the present invention;

[0059] Figure 3 is the two-dimensional energy spectrum of the second array before fusion of the present invention;

[0060] Figure 4 is the two-dimensional energy spectrum before the third array fusion of the present invention;

[0061] Figure 5 is the two-dimensional energy spectrum after the fusion of each array of the present invention;

[0062] Figure 6 is the positioning result of the target on the two-dimensional fusion energy spectrum of the present invention. Detailed implementation manners

[0063] The following further describes the present application in detail with reference to the accompanying drawings.

[0064] As Figure 1-4 shown, an embodiment of the present application discloses a cross-positioning method based on energy fusion and density clustering, including the following steps:

[0065] Step 1: Obtain the array received signal and generate a one-dimensional azimuth energy estimation spectrum. Among them, the array received signal is an active sonar signal, a passive sonar signal or an underwater communication signal. Specifically, the active sonar signal is an acoustic wave signal actively emitted by the sonar system. The acoustic wave propagates in water and is reflected back after encountering the target. The receiver realizes target detection and positioning by analyzing the echo signal; the passive sonar signal refers to the sonar system realizing target detection and positioning by receiving the acoustic signals radiated by the target itself (such as the mechanical noise and propeller noise of a submarine), without actively emitting acoustic waves; the underwater communication signal refers to an acoustic wave signal used for communication between underwater devices or between an underwater device and a surface device;

[0066] Step 2: Divide the region of interest into grids, and inversely map the one-dimensional azimuth energy estimation spectrum of each array to the two-dimensional grid region in combination with the propagation loss;

[0067] Step 3: Fuse the two-dimensional energy spectra of each array obtained in Step 2 to obtain a multi-array fusion energy spectrum on the two-dimensional grid region;

[0068] Step 4: Extract potential target points on the fusion energy spectrum by using an adaptive threshold technique according to the characteristics of the target points;

[0069] Step 5: Obtain the target position by using an adaptive density clustering method according to the cross-positioning characteristics and the characteristics of the potential target points. Among them, the density clustering method is the DBSCAN or OPTICS clustering algorithm, and other density-based clustering algorithms can also be used, which are not limited herein.

[0070] Specifically, the step of inversely mapping the one-dimensional azimuth energy estimation spectrum of each array to the two-dimensional grid region in combination with the propagation loss in Step 2 is:

[0071] Assume that the region of interest is The position of the center point of each array is ;

[0072] Calculate the angles and distances of each grid point relative to each array:

[0073]

[0074]

[0075] Among them, the angle is defined in the same way as the angle of the one-dimensional azimuth estimation spectrum, with due north as 0 degrees and increasing clockwise;

[0076] In the ideal case, the propagation loss from the grid point to each array, that is, calculate the propagation loss from the grid point to each array:

[0077]

[0078] Among them, is the propagation loss coefficient. In this example, the wavefront expands according to a spherical surface, that is , then calculate the two-dimensional energy spectrum of each array, and its formula is:

[0079]

[0080] Among them, is the one-dimensional azimuth energy estimation spectrum of the th array, is the grid point at the relative to the th array propagation loss coefficient.

[0081] Specifically, the fusion formula adopted in the fusion of the two-dimensional energy spectra of each array obtained in step three is:

[0082]

[0083] Among them, is the two-dimensional multi-array fusion energy spectrum, is the two-dimensional energy spectrum of the th array. When there is a target at the grid point, the energy inverted by each array to this grid point should be basically the same. After fusion, the energy of this grid point should be the minimum value in the local area of the fusion energy spectrum.

[0084] Specifically, the steps of extracting potential target points by using the adaptive threshold technology according to the characteristics of the target points on the fusion energy spectrum in step four are:

[0085] A1. Extract the ridge energy data in the fusion energy spectrum along the maximum angle in the one-dimensional azimuth estimation spectrum;

[0086] A2. Remove the ridge trend by polynomial fitting to obtain the detrended ridge data ;

[0087] A3. Combine the Pauta criterion and set the threshold to , where , to obtain potential target points that meet the conditions:

[0088] ;

[0089] Among them, when in an environment with high underwater noise, increase the value of to increase the threshold and reduce the false detection of noise points;

[0090] When in an environment with weak target signals, reduce the value of to lower the threshold and avoid missing the target.

[0091] Specifically, the steps of using the adaptive density clustering method to obtain the target position in step five are as follows:

[0092] B1. Set the minimum number of neighbors to M - 1, where M is the number of arrays;

[0093] B2. According to the main lobe width of each array at different angles, combined with the distance from the potential target point to the array, set different neighborhood distances for different potential target points:

[0094]

[0095] Among them, is the distance from the potential target point( to the center point of the array, is the angle formed by the potential target point and the array;

[0096] B3. Traverse all potential target points according to the density clustering method, and divide the potential target points with a distance less than the neighborhood distance into the same set. If the number of potential target points in the set is greater than the minimum number of neighbors, it is a positioning cluster, otherwise it is an interference cluster;

[0097] B4. According to the clusters obtained by clustering, the position of the positioning point is the mean value of the potential target points in each cluster:

[0098] .

[0099] Specifically, when the signal spreads in the form of a cylindrical wave, for example, when sound waves propagate in shallow water, the signal energy mainly diffuses in the horizontal direction. First, in the vertical direction, the energy attenuation is slow, and the energy attenuation is inversely proportional to the distance. At this time = 1;

[0100] When the signal expands in the form of a spherical wave, for example, when sound waves propagate in deep water, the signal energy diffuses uniformly in all directions, and the energy decays relatively fast. The energy decay is inversely proportional to the square of the distance. At this time = 2. In this example, the signal expands in the form of a spherical wave;

[0101] In addition, in the actual underwater environment, the propagation mode of sound waves may simultaneously include the characteristics of spherical waves and cylindrical waves, specifically depending on factors such as water depth, seabed topography, and sea surface conditions. The actual value of is between 1 and 2 and can be dynamically adjusted according to the actual scenario.

[0102] The following specifically describes this embodiment with reference to the accompanying drawings.

[0103] In this embodiment, the number of arrays is 3, and the coordinates of the array center positions are (-10 km, 0 km), (0 km, 0 km), and (10 km, 0 km). There are three targets in the region of interest. The position of target 1 is (-9 km, 15 km), the position of target 2 is (0 km, 20 km), and the position of target 3 is (13 km, 13 km). The two-dimensional energy spectra of each array obtained by combining the one-dimensional azimuth estimation spectrum and propagation loss inversion are as shown in the attachment Figure 2-4 shown. The two-dimensional fusion energy spectrum diagram obtained after fusion is as shown in Figure 5 shown. The final positioning result is as shown in Figure 6 shown. The obtained target point positions are (-8.97 km, 14.89 km), (0.00 km, 20.10 km), and (12.96 km, 12.85 km) respectively. The positioning errors are 0.11 km, 0.10 km, and 0.16 km, and there are no false detections or missed detections.

[0104] This method of cross-positioning based on energy fusion and density clustering effectively reduces the false target points in cross-positioning. The target positions obtained by this invention are close to the true positions, and the automatic extraction of target positions is realized, which has the advantages of convenience and easy operation.

[0105] The above are all the preferred embodiments of this application. The protection scope of this application is not limited accordingly. Therefore, all equivalent changes made according to the structure, shape, and principle of this application should be covered within the protection scope of this application.

Claims

1. A cross-location method based on energy fusion and density clustering, characterized in that, It includes the following steps: Step 1: Obtain the array received signal and generate a one-dimensional azimuth energy estimation spectrum; Step 2: Divide the region of interest into grids, and combine the propagation loss to invert the one-dimensional azimuth energy estimation spectrum of each array onto the two-dimensional grid region; Step 3: Fuse the two-dimensional energy spectra of each array obtained in Step 2 to obtain a multi-array fused energy spectrum on the two-dimensional grid region; Step 4: Extract potential target points on the fused energy spectrum using an adaptive threshold technique according to the characteristics of the target points; Step 5: Obtain the target position using an adaptive density clustering method based on the characteristics of cross-location and potential target points.

2. The cross - location method based on energy fusion and density clustering according to claim 1, wherein The step of combining the propagation loss to invert the one-dimensional azimuth energy estimation spectrum of each array onto the two-dimensional grid region in Step 2 is as follows: Assume that the region of interest is , and the positions of the centers of each array are ; Calculate the angles and distances of each grid point relative to each array: ； ； Calculate the propagation loss from the grid point to each array: ； Among them, is the propagation loss coefficient; Calculate the two-dimensional energy spectrum of each array: ； Among them, is the one-dimensional azimuth energy estimation spectrum of the th array, is the propagation loss coefficient at the grid point relative to the th array.

3. The cross-positioning method based on energy fusion and density clustering according to claim 1, characterized in that The fusion formula in Step 3 is: ； Among them, is a two-dimensional multi-array fusion energy spectrum, is the two-dimensional energy spectrum of the -th array.

4. A cross-location method based on energy fusion and density clustering according to claim 1, characterized in that, The step of extracting potential target points on the fused energy spectrum using an adaptive threshold technique according to the characteristics of the target points in Step 4 is as follows: A1. Extract the ridge energy data in the fused energy spectrum along the maximum value angle in the one-dimensional azimuth estimation spectrum; A2. Remove the ridge line trend by polynomial fitting to obtain the detrended ridge line data ; A3. Combine with the Pauta criterion and set the threshold to , where , to obtain potential target points that meet the conditions: 。 5. A cross-positioning method based on energy fusion and density clustering according to claim 1, characterized in that The step of obtaining the target position using an adaptive density clustering method in Step 5 is as follows: B1. Set the minimum number of neighborhoods to M - 1, where M is the number of arrays; B2. Set different neighborhood distances for different potential target points according to the main lobe widths at different angles of each array and the distance from the potential target point to the array; ； Among them, is the distance from the potential target point to the center point of the array, and is the angle formed by the potential target point and the array; B3. Traverse all potential target points according to the density clustering method, and divide the potential target points with distances less than the neighborhood distance into the same set. If the number of potential target points in the set is greater than the minimum number of neighborhoods, it is a positioning cluster, otherwise it is an interference cluster; B4. According to the clusters obtained by clustering, the position of the positioning point is the mean value of the potential target points in each cluster; 。 6. The cross - location method based on energy fusion and density clustering according to claim 1, characterized in that, The array received signal is an active sonar signal, a passive sonar signal, or an underwater communication signal.

7. A cross - location method based on energy fusion and density clustering according to claim 2, characterized in that, When the signal expands in the form of a cylindrical wave, the energy attenuation is inversely proportional to the distance, and in this case = 1; When the signal expands in the form of a spherical wave, the energy attenuation is inversely proportional to the square of the distance, and in this case = 2.

8. A cross-positioning method based on energy fusion and density clustering according to claim 4, characterized in that, When in an environment with high underwater noise, increase the value to increase the threshold and reduce the false detection of noise points; When in an environment with weak target signals, reduce the value to lower the threshold and avoid missing the detection of the target.

9. The cross - location method based on energy fusion and density clustering according to claim 1, characterized in that, In Step 5, the density clustering method is the DBSCAN or OPTICS clustering algorithm.

Images