Cluster-oriented special-shaped area multi-path search method and system

Through Delaunay triangulation method and K-means clustering algorithm, the special-shaped area is divided into detailed division, and combined with the plowing path generation method, the problem of uneven resource allocation and incomplete coverage in complex area searches is solved. , realizes efficient and regular area search and path planning.

CN118915819APending Publication Date: 2024-11-08ZHONGBING INTELLIGENT INNOVATION RES INST CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410971126.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-19
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

When existing drone clusters deal with complex or irregular geographical areas, it is difficult to accurately understand and divide work boundaries, resulting in waste of resources and low task execution efficiency. Especially in environments such as mountainous areas, forests or urban dense areas, the effect of collaborative operations is not Good, and there may be repeated coverage or missing areas.

Method used

Delaunay triangulation method and K-means clustering algorithm are used to initially divide and iteratively refine the search special-shaped areas to generate sub-regions with regular shapes and uniform areas. The search path is set using the plowing path generation method to ensure that no one is present efficiency and coverage integrity of machine parallel search.

Benefits of technology

Through regular sub-region division and adaptive path planning, the operation efficiency and coverage quality of the drone cluster are improved, resource waste and steering behavior frequency is reduced, effective coverage of complex areas is ensured, and task completion and stability are improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118915819A_ABST
    Figure CN118915819A_ABST
Patent Text Reader

Abstract

The invention relates to a cluster-oriented special-shaped region multi-path search method and system, belongs to the technical field of man-machine interaction, and solves the problem of non-uniform path planning and resource allocation when an existing cluster region is searched for an irregular special-shaped region. The method comprises the steps of obtaining a to-be-searched special-shaped area; and based on the number of unmanned platforms in a cluster, performing region division on the to-be-searched special-shaped region to obtain sub-regions with regular shapes and uniform areas, which comprises the following steps: performing initial sub-region division on the to-be-searched special-shaped region by using a Delaunay triangulation method, performing up-sampling in the to-be-searched special-shaped region to increase up-sampling points in the region, and performing up-sampling in the to-be-searched special-shaped region; performing fine division on each divided initial sub-region; and setting a search path for each sub-region, and performing parallel search by using an unmanned platform in the cluster. The irregular and special-shaped region is uniformly divided, so that unmanned platforms in a cluster can perform multi-machine collaborative search in parallel.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of human-computer interaction technology, and in particular to a cluster-oriented heterogeneous area multi-path search method and system. Background Art

[0002] In the frontier exploration of unmanned swarm technology, drone swarm systems are increasingly becoming a major focus in the field of intelligent technology. With their unique flexibility and scalability, these systems have demonstrated incomparable advantages in many application scenarios. In a broad sense, "unmanned swarms" aim to achieve automation and intelligent processing of complex tasks through a highly integrated intelligent system.

[0003] In existing cluster area search technologies, especially those involving drone applications, although existing technologies have demonstrated high efficiency and intelligence in many aspects, they still face a series of challenges and limitations. These technologies often expose shortcomings when dealing with complex or irregular geographical areas, especially in resource allocation and task execution efficiency. For example, in scenarios such as agriculture, search and rescue, and environmental monitoring, facing irregularly shaped areas, it is difficult for drone clusters to accurately understand and effectively divide their respective working boundaries, which not only leads to a waste of resources, but also limits the efficiency and quality of task execution.

[0004] For drone swarms, traditional methods often use preset paths or simple geometric divisions to allocate mission areas when facing complex terrain. This method is difficult to adapt to complex changes in terrain, especially in mountainous areas, forests, or densely populated urban areas. The collaborative operation effect between drones is greatly reduced. In addition, due to the lack of in-depth understanding of the internal structure of the region and the dynamic adjustment mechanism, drones may overlap or miss areas when performing coverage tasks, affecting the completion of the overall task. Summary of the invention

[0005] In view of the above analysis, the embodiments of the present invention aim to provide a cluster-oriented heterogeneous area multi-path search method and system to solve the problems of path planning and uneven resource allocation in the existing cluster area search in irregular heterogeneous areas.

[0006] The purpose of the present invention is mainly achieved through the following technical solutions:

[0007] The present invention provides a cluster-oriented heterogeneous area multi-path search method, comprising the following steps:

[0008] Get the alien-shaped area to be searched;

[0009] Based on the number of unmanned platforms in the cluster, the to-be-searched irregular-shaped area is divided into regions to obtain sub-regions with regular shapes and uniform areas, including: after the to-be-searched irregular-shaped area is divided into initial sub-regions using the Delaunay triangulation method, upsampling is performed in the to-be-searched irregular-shaped area to increase the upsampling points in the area, and each of the divided initial sub-regions is finely divided;

[0010] A search path is set for each of the sub-areas, and an unmanned platform in the cluster is used to perform parallel searches.

[0011] Furthermore, the initial region division of the to-be-searched irregular-shaped region using the Delaunay triangulation method includes:

[0012] Generate a super triangle; wherein the size of the super triangle can cover the special-shaped area to be searched;

[0013] Adding the vertices of the to-be-searched irregular-shaped area to the super triangle in sequence, and dividing the to-be-searched irregular-shaped area into a plurality of triangular areas using the Delaunay triangulation method;

[0014] The triangle centroids of the plurality of triangle areas are clustered based on the number of unmanned platforms in the cluster using a K-means algorithm to obtain a preset number of initial sub-areas; wherein the preset number is the number of unmanned platforms in the cluster.

[0015] Further, each of the initial sub-regions is iteratively refined to obtain a preset number of regional units; in one iteration:

[0016] Get the number of upsampling points and initial area units of this iteration;

[0017] Based on the number of up-sampling points, up-sampling is performed randomly in the heteromorphic area to be searched;

[0018] Based on the up-sampled points in each initial area unit, each initial area unit is triangulated to obtain a fine triangular mesh;

[0019] Using the K-means algorithm to cluster the centroids of the triangles in the fine triangular mesh based on a preset number, the irregular-shaped area to be searched is re-divided to obtain the area units of this iteration;

[0020] The number of up-sampling points in this iteration is multiplied by 2 as the number of up-sampling points in the next iteration, and each area unit in this iteration is used as each initial area unit in the next iteration, and the iteration is repeated;

[0021] When the shape and area of ​​each of the regional units meet the convergence condition, the iteration ends.

[0022] Furthermore, based on the sampling points in each initial area unit, each initial area unit is triangulated to obtain a fine triangular mesh, including: adding the up-sampling points in each initial area unit to each initial area unit in turn, and using the Delaunay triangulation method with the vertices of each initial area unit to divide the triangular area to obtain a fine triangular mesh.

[0023] Furthermore, during the first iteration:

[0024] Setting the number of sampling points based on the area of ​​the to-be-searched irregular-shaped area and the number of unmanned platforms in the cluster;

[0025] Each of the initial sub-regions serves as an initial region unit.

[0026] Further, setting a search path for each of the sub-regions includes:

[0027] Obtaining the principal axis of each of the sub-regions;

[0028] Based on the main axis of each sub-area, a cruising path for each sub-area is generated using a plowing path generation method;

[0029] Based on the length of each plowing line of the cruise path of each sub-area, the waypoints on each cruise path are set using an adaptive step size to obtain a search path for each sub-area.

[0030] Furthermore, obtaining the principal axis of each of the sub-regions includes:

[0031] Obtaining the centroid of each sub-region;

[0032] Perform a 180-degree bidirectional sweep with the centroid as the center of the circle, and extract the length of the boundaries of both sides of each sub-region on the sweep line every 1 degree;

[0033] The longest sweep line is selected as the main axis of each sub-region.

[0034] Furthermore, the generation of the cruising path of each sub-area using the plowing path generation method based on the main axis of each sub-area includes:

[0035] Taking the main axis of each sub-region as a reference line, generating a plurality of plowing lines parallel to the reference line at intervals of a preset width; wherein each plowing line takes the outermost edge of each sub-region as a boundary;

[0036] The plowing lines in each of the sub-areas are connected end to end to obtain the cruising path of each of the sub-areas.

[0037] Further, the method of using the adaptive step size to set the waypoints on each cruise path includes:

[0038] Set the baseline plowing line length and the baseline step length of the unmanned platform in the cluster;

[0039] Based on the length of each plowing line on each cruising path, the reference plowing line length and the reference step length, respectively calculating the step length of the unmanned platform for each plowing line on each cruising path;

[0040] Based on the step length of the unmanned platform on each plowing line, a waypoint of the unmanned platform is set on each plowing line.

[0041] The present invention also provides a cluster-oriented heterogeneous area multi-path search system, comprising:

[0042] A human-computer interaction module is used to obtain the special-shaped area to be searched;

[0043] A sub-region division module is used to divide the to-be-searched irregular-shaped region into sub-regions of regular shape and uniform area based on the number of unmanned platforms in the cluster; wherein, after the to-be-searched irregular-shaped region is initially divided into sub-regions using the Delaunay triangulation method, upsampling is performed in the to-be-searched irregular-shaped region to increase the upsampling points in the region, and each initial sub-region after division is finely divided;

[0044] The sub-area path planning module is used to set the search path for each sub-area and use the unmanned platforms in the cluster to perform parallel searches.

[0045] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects:

[0046] 1. The present invention supports user-defined search areas, divides the special-shaped area to be searched into sub-areas with regular shapes and equal areas, sets search paths for each sub-area according to its shape, and uses the unmanned platforms in the cluster to perform regional searches in parallel, thereby improving the operating efficiency of the unmanned platforms.

[0047] 2. The present invention integrates Delaunay triangulation technology, clustering algorithm and pyramid hierarchical theory when dividing the area, so that the shape of the segmented sub-areas is more regular, which is conducive to the simplification and optimization of path planning. The regular sub-areas not only ensure the uniformity after segmentation, avoiding the problem of uneven coverage caused by area differences, but also have clear boundaries with other sub-areas, improving the independence and efficiency of the coverage operation of each area, and providing a solid foundation for the efficient operation of the drone swarm.

[0048] 3. The present invention simplifies the computational complexity of the initial stage by first performing low-resolution segmentation on the searched heterogeneous area, can quickly find the initial solution, and avoids the local optimal solution trap.

[0049] 4. The present invention gradually refines the region division from low resolution to high resolution. The algorithm can gradually improve the accuracy and regularity of region division while maintaining computational efficiency. With the deepening of iteration, the algorithm's adaptability to complex terrain and shapes is gradually enhanced, ensuring the quality of the final region division.

[0050] 5. The path generation method in each sub-area of ​​the present invention further improves the coverage efficiency and integrity of the drone swarm through the selection of the main axis and adaptive step-length path generation, and overcomes the coverage omission or duplication problems that may exist in traditional static path planning.

[0051] 6. The path generation method of the present invention can greatly reduce the steering frequency of the UAV during the plowing process by extracting the main axis through sweeping, improve the stability of the UAV and reduce energy consumption. The sampling step size can be adaptively adjusted according to the size and shape of the plowing line, which improves the working efficiency of the UAV and ensures that the rugged working area can be effectively covered, thereby improving the efficiency and integrity of the operation.

[0052] In the present invention, the above-mentioned technical solutions can also be combined with each other to achieve more preferred combination solutions. Other features and advantages of the present invention will be described in the subsequent description, and some advantages can become obvious from the description, or can be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained through the contents particularly pointed out in the description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The drawings are only for the purpose of illustrating particular embodiments and are not to be considered limiting of the present invention. Like reference symbols denote like components throughout the drawings.

[0054] Figure 1 A schematic diagram of a flow chart of a cluster-oriented heterogeneous area multi-path search method in an embodiment of the present invention;

[0055] Figure 2 The diagram is a schematic diagram of the structure of a cluster-oriented heterogeneous area multi-path search system in an embodiment of the present invention. DETAILED DESCRIPTION

[0056] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0057] A specific embodiment of the present invention discloses a cluster-oriented heterogeneous area multi-path search method, such as Figure 1 As shown, the steps S1 to S3 are included:

[0058] Step S1, obtaining a special-shaped area to be searched.

[0059] Specifically, the user manually draws a target area to be searched in a given map. The target area can be a polygon of any shape without restrictions on the number of sides or side lengths.

[0060] Step S2, based on the number of unmanned platforms in the cluster, the to-be-searched irregular-shaped area is divided into regions to obtain sub-regions with regular shapes and uniform areas, including: using the Delaunay triangulation method to perform initial sub-region division on the to-be-searched irregular-shaped area; performing upsampling in the to-be-searched irregular-shaped area to increase the upsampling points in the area, and performing fine division on each of the divided initial sub-regions.

[0061] Specifically, based on the number of unmanned platforms in the cluster, the to-be-searched irregular-shaped area is divided into sub-areas, the divided sub-areas correspond to the unmanned platforms one by one, and the unmanned platforms are used to perform parallel searches on the corresponding sub-areas in the to-be-searched irregular-shaped area to improve the search efficiency; the to-be-searched irregular-shaped area is divided into sub-areas with regular shapes and uniform areas, so that the unmanned platforms in the cluster can avoid waste of resources and search blind spots, simplify path planning, plan search paths more efficiently, reduce the frequency of turning behaviors, improve stability and reduce energy consumption.

[0062] Furthermore, the initial region division of the to-be-searched irregular region using the Delaunay triangulation method includes steps S201 to S203:

[0063] Step S201, generating a super triangle; wherein the size of the super triangle can cover the special-shaped area to be searched.

[0064] Specifically, the super triangle is a triangle that is large enough to ensure that the irregular area to be searched is located inside the super triangle. In the initial stage of Delaunay triangulation, the super triangle is used as the basic structure to gradually insert the vertices of the irregular area into the triangle to construct a triangulated network of the entire area.

[0065] Step S202: adding the vertices of the to-be-searched irregular-shaped area to the super triangle in sequence, and dividing the to-be-searched irregular-shaped area into a plurality of triangular areas using the Delaunay triangulation method.

[0066] Specifically, each vertex of the to-be-searched irregular area is gradually added to the super triangle. After each vertex is inserted, the structure of the triangulated network is dynamically adjusted to meet the Delaunay condition, that is, the circumscribed circle of any triangle does not contain any other vertices and maximizes the minimum angle characteristic. These two characteristics avoid the generation of narrow and long triangles and make the divided triangles more regular.

[0067] Step S203, clustering the triangle centroids of the plurality of triangle areas based on the number of unmanned platforms in the cluster using a K-means algorithm to obtain a preset number of initial sub-areas; wherein the preset number is the number of unmanned platforms in the cluster.

[0068] Specifically, for several triangular areas divided into the heterogeneous area to be searched, the K-means algorithm is used to cluster the centroids of the triangles, and the number of clusters K is the number of unmanned platforms in the cluster. The purpose is to preliminarily divide the heterogeneous area to be searched into K independent initial sub-areas. In the initial division stage, only the vertices of the heterogeneous area to be searched are used for division and clustering, which can greatly reduce the amount of calculation and quickly obtain a rough division result. Upsampling and refined triangulation are performed on the results of the initial division, which can avoid the area division from falling into a local optimal solution and help to balance the overall area division.

[0069] Furthermore, each of the initial sub-regions is iteratively refined to obtain a preset number of region units with regular shapes and equal areas.

[0070] Specifically, the irregularly shaped area to be searched is upsampled to increase the density of points in the area, thereby performing more precise triangulation to adjust the triangle boundary, ensuring that the shape of each sub-area is more regular and the area distribution is more balanced.

[0071] Furthermore, in one iteration process, the following steps S211 to S216 are included:

[0072] Step S211, obtaining the number of up-sampling points and each initial area unit of this iteration.

[0073] Furthermore, in the first iteration process, the number of up-sampling points is set based on the area of ​​the to-be-searched irregular-shaped area and the number of unmanned platforms in the cluster; each of the initial sub-areas is used as each initial area unit.

[0074] Specifically, the number of sampling points is set based on the area of ​​the to-be-searched irregular-shaped area and the number of unmanned platforms in the cluster, including:

[0075] Divide the total area of ​​the irregular-shaped region to be searched by the number of regions to be divided to obtain the ideal area of ​​each sub-region, that is, divide the area of ​​the irregular-shaped region to be searched in half;

[0076] The area of ​​each triangle of the to-be-searched irregular-shaped area divided in step S202 is calculated respectively, wherein the area of ​​the triangle is calculated using Heron's formula:

[0077]

[0078] Among them, S i represents the area of ​​the i-th triangle; x i ,y i 、z i are the lengths of the three sides of the i-th triangle respectively;

[0079] Sum the areas of the triangles in each initial sub-region to obtain the area of ​​each initial sub-region;

[0080] Based on the area of ​​each initial sub-region and the ideal sub-region, the number of upsampling points is obtained using the following formula:

[0081]

[0082] Where A represents the number of upsampling points; K represents the number of initial sub-regions; S j represents the area of ​​the jth initial sub-region; S ideal represents the ideal sub-region area; P represents the number of preset reference sampling points. Exemplarily, the number of reference sampling points can be set to 100.

[0083] Step S212: Based on the number of up-sampling points, up-sampling is performed randomly in the to-be-searched irregular region.

[0084] Specifically, sampling is performed by randomly scattering points in the irregular area to be searched, which can adapt to irregular areas, achieve uniform coverage in the area, and reduce coverage blind spots and overlapping areas.

[0085] Step S213: triangulate each initial area unit based on the up-sampled points in each initial area unit to obtain a fine triangular mesh.

[0086] Furthermore, the up-sampled points in each initial area unit are added to each initial area unit in turn, and the Delaunay triangulation method is used with the vertices of each area unit to divide the triangular area to obtain a fine triangular mesh.

[0087] Specifically, the up-sampling points in each initial area unit are sequentially inserted into each current initial area unit, and each time an up-sampling point is inserted, the triangulation structure is dynamically adjusted to meet the Delaunay condition.

[0088] Step S214: clustering the centroids of the triangles in the fine triangular mesh based on a preset number using a K-means algorithm, and re-dividing the to-be-searched irregular-shaped area to obtain area units of this iteration.

[0089] Specifically, in the divided irregular area to be searched, the K-means clustering algorithm is used to cluster the centroids of each triangle in the fine triangular mesh again, and it is redivided into K area units to ensure that the shape of each area is more regular and the area distribution is more balanced.

[0090] Step S215, multiplying the number of up-sampling points in this iteration by 2 as the number of up-sampling points in the next iteration, taking each area unit in this iteration as each initial area unit in the next iteration, and iterating repeatedly;

[0091] Specifically, in each iteration process, the sampling points in each initial area unit are gradually increased to perform a finer triangulation.

[0092] Step S216: When the shape and area of ​​each of the regional units meet the convergence condition, the iteration ends.

[0093] Specifically, the convergence condition is that the regional units have regular shapes and equal areas.

[0094] It should be noted that after all the division steps are completed, all triangles containing the vertices of the super triangle are removed to complete all the steps of Delaunay triangulation.

[0095] Step S3: setting a search path for each of the sub-areas, and using the unmanned platforms in the cluster to perform parallel searches.

[0096] Further, the method includes the following steps S301 to S303:

[0097] Step S301, obtaining the principal axis of each sub-region, comprising:

[0098] Obtaining the centroid of each sub-region;

[0099] Perform a 180-degree bidirectional sweep with the centroid as the center of the circle, and extract the length of the boundaries of both sides of each sub-region on the sweep line every 1 degree;

[0100] The longest sweep line is selected as the main axis of each sub-region.

[0101] Specifically, by extracting the main axis of each sub-area and planning a path along its direction, the unmanned platform in the cluster can reduce the frequency of turning behavior when performing tasks, which can not only improve the stability of the unmanned platform, but also reduce energy consumption and extend the operation time of the unmanned platform.

[0102] Step S302: Based on the main axis of each sub-area, a cruising path for each sub-area is generated using a plowing path generation method.

[0103] Furthermore, generating a cruise path for each of the sub-areas includes:

[0104] Taking the main axis of each sub-region as a reference line, a plurality of plowing lines parallel to the reference line are generated at intervals of a preset width; wherein each plowing line takes the outermost edge of each sub-region as a boundary.

[0105] Specifically, the starting point and the end point of each plowing line in each sub-area are located at the edge of each sub-area, and the preset width between each plowing line can be set as the scanning range of the unmanned platform, for example.

[0106] The plowing lines in each of the sub-areas are connected end to end to obtain the cruising path of each of the sub-areas.

[0107] Specifically, taking any boundary point of the farthest plowing line on the side perpendicular to the main axis of each sub-area as the starting point, the other boundary point is used as the tail of the current plowing line and connected with the boundary point of the adjacent plowing line along the edge line of the sub-area, so that each plowing line forms a continuous path with the adjacent plowing line until all plowing lines are connected to form a continuous path, which serves as the cruising path of the unmanned platform of each sub-area.

[0108] Step S303: based on the lengths of the plowing lines of the cruise paths of the sub-areas, the waypoints on the cruise paths are set using an adaptive step size to obtain a search path for each sub-area.

[0109] Further, the method of using the adaptive step size to set the waypoints on each cruise path includes:

[0110] Set the baseline plowing line length and the baseline step length of the unmanned platforms in the cluster.

[0111] Specifically, the average of the plowing line lengths in each of the sub-areas is taken as the benchmark plowing line length, and the benchmark plowing line length is evenly divided to obtain the benchmark step length of the unmanned platform. It should be noted that the benchmark step length is smaller than the maximum scanning range of the unmanned platform.

[0112] Based on the length of each plowing line on each cruising path, the reference plowing line length and the reference step length, the unmanned platform step length of each plowing line on each cruising path is calculated respectively.

[0113] Specifically, the step length of the unmanned platform for each plowing line is obtained using the following formula:

[0114]

[0115] Among them, STRI m represents the step length of the unmanned platform on the mth plowing line; STRI base Indicates the benchmark step length; L m represents the length of the mth plowing line; Lbase Indicates the length of the reference plowing line.

[0116] It should be noted that when the calculated unmanned platform step length of the plowing line is smaller than the minimum step length of the unmanned platform, the minimum step length of the unmanned platform is set as the unmanned platform step length of the current plowing line.

[0117] Based on the step length of the unmanned platform on each plowing line, a waypoint of the unmanned platform is set on each plowing line.

[0118] Specifically, the waypoint is a preset geographic coordinate point for the unmanned platform. The unmanned platform will navigate according to these points when performing the search mission. In this embodiment, two boundary points of each plowing line are set as waypoints. Starting from a boundary point, every STRI m A waypoint is set on the plowing line until the other boundary point on this plowing line. The step length between plowing paths is dynamically adjusted according to the length of different plowing lines. A larger step length is used on a longer plowing line to speed up the coverage speed, while the step length is automatically reduced for a shorter or complex plowing line, ensuring that the unmanned platform can achieve the best coverage effect in paths of different lengths while avoiding missing corners.

[0119] Another embodiment of the present invention discloses a cluster-oriented heterogeneous area multi-path search system, such as Figure 2 As shown, including:

[0120] A human-computer interaction module is used to obtain the special-shaped area to be searched;

[0121] A sub-region division module is used to divide the to-be-searched irregular-shaped region into sub-regions of regular shape and uniform area based on the number of unmanned platforms in the cluster; wherein, after the to-be-searched irregular-shaped region is initially divided into sub-regions using the Delaunay triangulation method, upsampling is performed in the to-be-searched irregular-shaped region to increase the upsampling points in the region, and each initial sub-region after division is finely divided;

[0122] The sub-area path planning module is used to set the search path for each sub-area and use the unmanned platforms in the cluster to perform parallel searches.

[0123] The above system embodiment and method embodiment are based on the same principle, and their related parts can be used for reference, and can achieve the same technical effect. The specific implementation process refers to the above embodiment, which will not be repeated here.

[0124] In summary, the cluster-oriented heterogeneous area multi-path search method and system according to the embodiment of the present invention have the following beneficial effects:

[0125] 1. The present invention supports user-defined search areas, divides the special-shaped area to be searched into sub-areas with regular shapes and equal areas, sets search paths for each sub-area according to its shape, and uses the unmanned platforms in the cluster to perform regional searches in parallel, thereby improving the operating efficiency of the unmanned platforms.

[0126] 2. The present invention integrates Delaunay triangulation technology, clustering algorithm and pyramid hierarchical theory when dividing the area, so that the shape of the segmented sub-areas is more regular, which is conducive to the simplification and optimization of path planning. The regular sub-areas not only ensure the uniformity after segmentation, avoiding the problem of uneven coverage caused by area differences, but also have clear boundaries with other sub-areas, improving the independence and efficiency of the coverage operation of each area, and providing a solid foundation for the efficient operation of the drone swarm.

[0127] 3. The present invention simplifies the computational complexity of the initial stage by first performing low-resolution segmentation on the searched heterogeneous area, can quickly find the initial solution, and avoids the local optimal solution trap.

[0128] 4. The present invention gradually refines the region division from low resolution to high resolution. The algorithm can gradually improve the accuracy and regularity of region division while maintaining computational efficiency. With the deepening of iteration, the algorithm's adaptability to complex terrain and shapes is gradually enhanced, ensuring the quality of the final region division.

[0129] 5. The path generation method in each sub-area of ​​the present invention further improves the coverage efficiency and integrity of the drone swarm through the selection of the main axis and adaptive step-length path generation, and overcomes the coverage omission or duplication problems that may exist in traditional static path planning.

[0130] 6. The path generation method of the present invention can greatly reduce the steering frequency of the UAV during the plowing process by extracting the main axis through sweeping, improve the stability of the UAV and reduce energy consumption. The sampling step size can be adaptively adjusted according to the size and shape of the plowing line, which improves the working efficiency of the UAV and ensures that the rugged working area can be effectively covered, thereby improving the efficiency and integrity of the operation.

[0131] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A cluster-oriented heterogeneous area multi-path search method, characterized in that: The steps include: Get the alien-shaped area to be searched; Based on the number of unmanned platforms in the cluster, the to-be-searched irregular-shaped area is divided into regions to obtain sub-regions with regular shapes and uniform areas, including: after the to-be-searched irregular-shaped area is divided into initial sub-regions using the Delaunay triangulation method, upsampling is performed in the to-be-searched irregular-shaped area to increase the upsampling points in the area, and each of the divided initial sub-regions is finely divided; A search path is set for each of the sub-areas, and an unmanned platform in the cluster is used to perform parallel searches.

2. The method according to claim 1, characterized in that: The initial region division of the to-be-searched irregular region using the Delaunay triangulation method comprises: Generate a super triangle; wherein the size of the super triangle can cover the special-shaped area to be searched; Adding the vertices of the to-be-searched irregular-shaped area to the super triangle in sequence, and dividing the to-be-searched irregular-shaped area into a plurality of triangular areas using the Delaunay triangulation method; The triangle centroids of the plurality of triangle areas are clustered based on the number of unmanned platforms in the cluster using a K-means algorithm to obtain a preset number of initial sub-areas; wherein the preset number is the number of unmanned platforms in the cluster.

3. The method according to claim 2, characterized in that: Iteratively refine each of the initial sub-regions to obtain a preset number of regional units; in one iteration: Get the number of upsampling points and initial area units of this iteration; Based on the number of up-sampling points, up-sampling is performed randomly in the heteromorphic area to be searched; Based on the up-sampled points in each initial area unit, each initial area unit is triangulated to obtain a fine triangular mesh; Using the K-means algorithm to cluster the centroids of the triangles in the fine triangular mesh based on a preset number, the irregularly shaped area to be searched is re-divided to obtain the area units of this iteration; The number of up-sampling points in this iteration is multiplied by 2 as the number of up-sampling points in the next iteration, and each area unit in this iteration is used as each initial area unit in the next iteration, and the iteration is repeated; When the shape and area of ​​each of the regional units meet the convergence condition, the iteration ends.

4. The method according to claim 3, characterized in that: Based on the sampling points in each initial area unit, each initial area unit is triangulated to obtain a fine triangular mesh, including: adding the up-sampling points in each initial area unit to each initial area unit in turn, and using the Delaunay triangulation method with the vertices of each initial area unit to divide the triangular area to obtain a fine triangular mesh.

5. The method according to claim 3, characterized in that: During the first iteration: Setting the number of sampling points based on the area of ​​the to-be-searched irregular-shaped area and the number of unmanned platforms in the cluster; Each of the initial sub-regions serves as an initial region unit.

6. The method according to claim 1, characterized in that: Setting a search path for each of the sub-areas includes: Obtaining the principal axis of each of the sub-regions; Based on the main axis of each sub-area, a cruising path for each sub-area is generated using a plowing path generation method; Based on the length of each plowing line of the cruise path of each sub-area, the waypoints on each cruise path are set using an adaptive step size to obtain a search path for each sub-area.

7. The method according to claim 6, characterized in that: The obtaining of the principal axis of each sub-region comprises: Obtaining the centroid of each sub-region; Perform a 180-degree bidirectional sweep with the centroid as the center of the circle, and extract the length of the boundaries of both sides of each sub-region on the sweep line every 1 degree; The longest sweep line is selected as the main axis of each sub-region.

8. The method according to claim 7, characterized in that: The generating of the cruising path of each sub-area by using the plowing path generating method based on the main axis of each sub-area comprises: Taking the main axis of each sub-region as a reference line, generating a plurality of plowing lines parallel to the reference line at intervals of a preset width; wherein each plowing line takes the outermost edge of each sub-region as a boundary; The plowing lines in each of the sub-areas are connected end to end to obtain the cruising path of each of the sub-areas.

9. The method according to claim 8, characterized in that: The method of using the adaptive step size to set the waypoints on each cruise path includes: Set the baseline plowing line length and the baseline step length of the unmanned platform in the cluster; Based on the length of each plowing line on each cruising path, the reference plowing line length and the reference step length, respectively calculating the step length of the unmanned platform for each plowing line on each cruising path; Based on the step length of the unmanned platform on each plowing line, a waypoint of the unmanned platform is set on each plowing line.

10. A cluster-oriented heterogeneous area multi-path search system, characterized in that: include: A human-computer interaction module is used to obtain the special-shaped area to be searched; A sub-region division module is used to divide the to-be-searched irregular-shaped region into sub-regions of regular shape and uniform area based on the number of unmanned platforms in the cluster; wherein, after the to-be-searched irregular-shaped region is initially divided into sub-regions using the Delaunay triangulation method, upsampling is performed in the to-be-searched irregular-shaped region to increase the upsampling points in the region, and each initial sub-region after division is finely divided; The sub-area path planning module is used to set the search path for each sub-area and use the unmanned platforms in the cluster to perform parallel searches.