A network graph partitioning method and system for multi-UAV line patrol

Abstract

The present invention discloses a network graph partitioning method and system for multi-UAV line patrol, the method comprising: generating a task network graph for UAV patrol; partitioning the task network graph by a first algorithm to obtain a first partition set; screening the first partition set by a second algorithm to obtain a feasible solution set; calculating the feasible solution set to obtain a network graph partitioning scheme for the UAV line patrol. The present invention adopts the basic cut set principle to partition the network graph, has strong practicability, and generates a high-quality partitioning scheme.

Description

Technical Field

[0001] The present invention relates to the technical field of multi-UAV line patrol task allocation, and in particular to a network graph partitioning method and system for multi-UAV line patrol. Background Art

[0002] UAVs are involved in network patrol in applications such as road network patrol, power grid inspection and oil pipeline inspection. Faced with increasingly complex environments and demands, it is often necessary to consider multiple UAVs to complete tasks. Therefore, the allocation of multi-UAV patrol tasks has become a hot topic for many scholars. The multi-UAV patrol task allocation problem is essentially a network graph partition problem. The existing common technical solutions are mainly geometric partitioning, edge k clustering and some heuristic algorithms.

[0003] The geometric partitioning method calculates the partitioning scheme based on the coordinate information of the nodes in the grid without considering the connectivity between the nodes. Although this method has a fast partitioning speed, the quality of the generated partitioning scheme is poor; edge k clustering is based on the number of drones k, and each edge is divided into k areas based on the principle of geographical proximity, and the target weights in each area are guaranteed to be roughly equal. These k areas are the patrol areas assigned to k drones, but the connectivity of the divided areas is poor, which increases the planning complexity; most heuristic algorithms refer to computational improvement strategies designed for a specific problem, but they are highly dependent on actual problems and have low scenario scalability. Summary of the invention

[0004] In view of this, the technical problem solved by the present invention is to provide a network graph partitioning method and system for multi-UAV line patrol, which adopts the basic cut set principle to divide the network graph, has strong practicality, and generates a high-quality partitioning scheme.

[0005] The present invention discloses a network graph partitioning method for multi-UAV line patrol, which comprises:

[0006] Generate a mission network diagram for drone patrols;

[0007] Partition the task network graph using a first algorithm to obtain a first partition set;

[0008] Screening the first partition set by a second algorithm to obtain a feasible solution set;

[0009] The feasible solution set is calculated to obtain a network graph partitioning scheme for the drone line patrol.

[0010] Preferably, the step of generating a task network diagram for drone patrol includes:

[0011] Assume that the connected graph G =<V,E,W> is the task network diagram of p (p ≥ 2) drone patrols, let v 0 ,v1 ,…,v m ∈V represents the point set, e 0 ,e 1 ,…,e n ∈E represents the edge set, w 0 ,w 1 ,…,w n ∈W represents the weight set of the edge, where w i It is edge i The weights can express the navigation parameters.

[0012] Preferably, the navigation parameters include navigation distance and navigation income.

[0013] Preferably, the task network graph is divided into p subgraphs as the task areas of the UAVs, so that the target value of each UAV is as balanced as possible; assuming that the single-flight capability of the UAVs meets the task volume of the sub-area, the navigation trajectory is an Euler circuit, with the goal of balancing the navigation distance.

[0014] Preferably, the task network diagram is partitioned by a first algorithm to obtain a first partition set, including:

[0015] S51: Obtain a spanning tree T of the task network graph G;

[0016] S52: Generate '2-partition' set D= <d 1 ,d 2 ,...>;

[0017] S53: traverse all branches of tree T, a' j and a″ j Represents branches S j The two endpoints of

[0018] S54: Get TS j Center and point a' j Connected point set V′ j , TS j Center and point a″ j Connected point set V″ j ;

[0019] S55: Divide G into three graphs, Figure 1 Remember G' j = <V′ j ,E' j ,W′ j >, sub Figure 2 Remember G″ j = <V″ j ,E″ j ,W″ j >, cut graph G″′j = <V″′ j ,E″′ j ,W″′ j >, satisfying V=V′ j ∪V″ j , E=E' j ∪E″ j ∪E″′ j ,further

[0020] S56: E″′ j The edges in are randomly assigned to E' j and E″ j , and get d j = <G' j ,G″ j >, that is, the first partition set is obtained.

[0021] Preferably, the step of screening the first partition set by a second algorithm to obtain a feasible solution set comprises the following steps:

[0022] S61: Set the equilibrium threshold α and the number of drones p (p≥2);

[0023] S62: traverse the partitioning scheme d in D j ;

[0024] S63: Find d respectively j The Euler graph of the neutron subgraph and the edge weight sum of the Euler subgraph are calculated, denoted by dw j = <gw' j ,gw″ j >

[0025] S64: Retain the content (1-α)·p -1 ≤gw' j ·(gw' j +gw″ j ) -1 ≤(1+α)·p -1 ;

[0026] S65: If p-1=1, output the solution retained in step S64, and then enter the next group d in step S62. j ; Otherwise, call the first algorithm for G″ j Divide to obtain the second partition set and traverse, and then go to step S63; wherein, record g p Represents the graph structure divided by the p-th drone;

[0027] S66: Record the selected feasible solutions. The feasible solution set is obtained.

[0028] Preferably, the first partition set is a 2-partition set under a tree of the task network graph.

[0029] Preferably, the calculating of the feasible solution set to obtain the network graph partitioning scheme for the drone line patrol includes:

[0030] The variance of the target value of each feasible solution in the feasible solution set is calculated, and the feasible solution corresponding to the minimum variance in the variance is taken as the optimal solution output. The optimal solution is the network graph partitioning scheme for the drone line patrol.

[0031] The present invention also discloses a network graph partitioning system for multi-UAV line patrol, comprising:

[0032] A generation module is used to generate a task network diagram for drone patrol;

[0033] A first calculation module, configured to partition the task network graph using a first algorithm to obtain a first partition set;

[0034] A second computing module, configured to screen the first partition set using a second algorithm to obtain a feasible solution set;

[0035] The third calculation module is used to calculate the feasible solution set to obtain the network graph partitioning scheme for the drone line patrol.

[0036] Due to the adoption of the above technical solution, the present invention has the following advantages: (1) The basic cut set principle is used to divide the network graph, which is highly practical and the generated division scheme is of high quality; (2) The method based on the present invention can construct multiple target solutions, which is helpful for solving the optimization model. (3) The task allocation of network line patrol of two or more drones can be quickly realized, and the scheme can be flexibly adjusted according to user needs, which has certain value in engineering applications; (4) The application scenario of the present invention can not only be used for multi-drone line patrol, but also for multi-unmanned vehicle / manned vehicle line patrol. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the embodiments of the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0038] Figure 1 A schematic diagram of a flow chart of a method for dividing a network graph of a multi-UAV line patrol according to an embodiment of the present invention;

[0039] Figure 2A schematic diagram of '2-partitioning' of a task network diagram according to an embodiment of the present invention;

[0040] Figure 3 A schematic diagram of a multi-UAV task allocation solution according to an embodiment of the present invention;

[0041] Figure 4 A schematic diagram of the structure of a network graph partitioning system for multi-UAV line patrol according to an embodiment of the present invention. DETAILED DESCRIPTION

[0042] The present invention is further described in conjunction with the accompanying drawings and embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by ordinary technicians in this field should fall within the scope of protection of the embodiments of the present invention.

[0043] See also Figure 1 The present invention provides an embodiment of a network graph partitioning method for multi-UAV line patrol, the method comprising the following steps:

[0044] S101: Generate a task network diagram for drone patrol;

[0045] S102: Divide the task network graph by a first algorithm to obtain a first partition set;

[0046] S103: Screening the first partition set by using the second algorithm to obtain a feasible solution set;

[0047] S104: Calculate the feasible solution set to obtain a network graph partitioning scheme for drone line patrol.

[0048] In this embodiment, generating a task network diagram for drone patrol includes:

[0049] Assume that the connected graph G =<V,E,W> is the task network diagram of p (p ≥ 2) drone patrols, let v 0 ,v 1 ,…,v m ∈V represents the point set, e 0 ,e 1 ,…,e n ∈E represents the edge set, w 0 ,w 1 ,…,w n ∈W represents the weight set of the edge, where w i It is edge i The weights can express the navigation parameters.

[0050] In this embodiment, the navigation parameters include navigation distance and navigation income.

[0051] In this embodiment, the task network graph is divided into p subgraphs as the task areas of the UAVs, so that the target value of each UAV is as balanced as possible; it is assumed that the single-flight capability of the UAVs meets the task volume of the sub-area, and the navigation trajectory is an Euler circuit, with the goal of balancing the navigation distance.

[0052] In this embodiment, the task network graph is divided by a first algorithm to obtain a first partition set, including:

[0053] S51: Obtain a spanning tree T of the task network graph G;

[0054] S52: Generate '2-partition' set D= <d 1 ,d 2 ,...>;

[0055] S53: traverse all branches of tree T, a' j and a″ j Represents branches S j The two endpoints of

[0056] S54: Get TS j Center and point a' j Connected point set V′ j , TS j Center and point a″ j Connected point set V″ j ;

[0057] S55: Divide G into three graphs, Figure 1 Remember G' j = <V′ j ,E' j ,W′ j >, sub Figure 2 Remember G″ j = <V″ j ,E″ j ,W″ j >, cut graph G″′ j = <V″ j ,E″′ j ,W″′ j >, satisfying V=V′ j ∪V″ j , E=E' j ∪E″ j ∪E″′ j ,further

[0058] S56: E″′ j The edges in are randomly assigned to E' j and E″′ j , and get d j = <G'j ,G″′ j >, that is, the first partition set is obtained.

[0059] In this embodiment, the first partition set is screened by the second algorithm to obtain a feasible solution set, including the following steps:

[0060] S61: Set the equilibrium threshold α and the number of drones p (p≥2);

[0061] S62: traverse the partitioning scheme d in D j ;

[0062] S63: Find d respectively j The Euler graph of the neutron subgraph and the edge weight sum of the Euler subgraph are calculated, denoted by dw j = <gw' j ,gw″′ j >

[0063] S64: Retain the content (1-α)·p -1 ≤gw' j ·(gw' j +gw″ j ) -1 ≤(1+α)·p -1 ;

[0064] S65: If p-1=1, then output the solution retained in step S64, and then enter the next group d in step S62 j ; Otherwise, call the first algorithm for G″ j Divide to obtain the second partition set and traverse, and go to step S63; wherein, record g p Represents the graph structure divided by the p-th drone;

[0065] S66: Record the selected feasible solutions. Get a feasible solution set.

[0066] In this embodiment, the first partition set is a 2-partition set under a tree of the task network graph.

[0067] In this embodiment, the feasible solution set is calculated to obtain a network graph partitioning scheme for drone line patrol, including:

[0068] The variance of the target value of each feasible solution in the feasible solution set is calculated, and the feasible solution corresponding to the minimum variance in the variance is taken as the optimal solution output. The optimal solution is the network graph partitioning scheme for drone line inspection.

[0069] See also Figure 2 and Figure 3,This invention also gives a specific case for process analysis and ,demonstration. The algorithm is written and run in the python 3.9 ,environment.

[0070] See also Figure 2 , generate a task network example diagram as required Figure 2 As shown in (1), calling the first algorithm, the resulting spanning tree is as follows Figure 2 As shown in (2), Figure 2 The division results corresponding to the two branches are listed in Figure 2 (3)- Figure 2 In (5), the tree is divided according to branches (4,6). Figure 2 (4) The dashed edge in the middle is the cut edge. The two sides of the cut edge are two subgraphs. Then the cut edges are evenly distributed to obtain the "2-partition" result. Figure 2 (5) Similarly, Figure 2 (6)- Figure 2 (8) is the result of '2-partition' of another branch (6,13), that is, Figure 2 Some results for the first partition set in this specific case are given.

[0071] See also Figure 3 , the test results of the second algorithm are as follows Figure 3 As shown in the figure, the feasible line patrol task allocation scheme when the number of drones is 3, 4, 5 and 6, that is, the network graph partitioning scheme of drones, is shown.

[0072] See also Figure 4 The present invention also discloses an embodiment of a network graph partitioning system for multi-UAV line patrol, the system comprising:

[0073] A generation module 401 is used to generate a task network diagram for drone patrol;

[0074] A first calculation module 402 is used to partition the task network graph using a first algorithm to obtain a first partition set;

[0075] A second calculation module 403 is used to screen the first partition set by a second algorithm to obtain a feasible solution set;

[0076] The third calculation module 404 is used to calculate the feasible solution set to obtain a network graph partitioning scheme for drone line patrol.

[0077] The system embodiment described above is merely illustrative, wherein the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, i.e., they may be located in one place, or they may be distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative effort.

[0078] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, or of course by hardware. Based on such an understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, and the computer-readable recording medium includes any mechanism for storing or transmitting information in a form readable by a computer (e.g., a computer).

[0079] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A network graph partitioning method for multi-UAV line inspection, It is characterized in that include: Generate a mission network diagram for drone patrols; Partition the task network graph using a first algorithm to obtain a first partition set; Screening the first partition set by a second algorithm to obtain a feasible solution set; Calculate the feasible solution set to obtain a network graph partitioning scheme for the drone line patrol; The task network diagram for generating a drone patrol includes: Assume that the connected graph G =<V,E,W> is the task network diagram of p (p ≥ 2) drone patrols, let v 0 ,v 1 ,…,v m ∈V represents the point set, e 0 ,e 1 ,…,e n ∈E represents the edge set, w 0 ,w 1 ,…,w n ∈W represents the weight set of the edge, where w i It is edge i The weights of , which express the navigation parameters; The step of partitioning the task network graph by a first algorithm to obtain a first partition set includes: S51: Obtain a spanning tree T of the task network graph G; S52: Generate '2-partition' set D= <d 1 ,d 2 ,...>; S53: traverse all branches of tree T, a' j and a″ j Represents branches S j The two endpoints of S54: Get TS j Center and point a' j Connected point set V j ', TS j Center and point a″ j Connected point set V″ j ; S55: Divide G into three graphs, subgraph 1 is called G' j = <V j ',E' j ,W j '>, subgraph 2 is G' j '= <V j ”,E' j ',W j ">, cut graph G' j ”= <V j ”',E' j ”,W j ”'>, satisfying V=V j '∪V j ”, E=E' j ∪E″ j ∪E″′ j ,further S56: E' j The edges in " are randomly assigned to E' j and E' j ', get d j = <G' j ,G' j '>, that is, the first partition set is obtained; The method of screening the first partition set by using a second algorithm to obtain a feasible solution set comprises the following steps: S61: Set the equilibrium threshold α and the number of drones p (p≥2); S62: traverse the partitioning scheme d in D j ; S63: Find d respectively j The Euler graph of the neutron subgraph and the edge weight sum of the Euler subgraph are calculated, denoted by dw j = <gw' j ,gw″ j > S64: Retain the content (1-α)·p -1 ≤gw' j ·(gw' j +gw″ j ) -1 ≤(1+α)·p -1 ; S65: If p-1=1, output the solution retained in step S64, and then enter the next group d in step S62. j ; Otherwise, call the first algorithm for G″ j Divide to obtain the second partition set and traverse, and then go to step S63; wherein, record g p Represents the graph structure divided by the p-th drone; S66: Record the feasible solutions selected. The feasible solution set is obtained.

2. The method according to claim 1, It is characterized in that The navigation parameters include navigation distance and navigation income.

3. The method according to claim 1, It is characterized in that The task network graph is divided into p subgraphs as the task areas of the UAV, so that the target value of each UAV is as balanced as possible; assuming that the single-flight capability of the UAV meets the task volume of the sub-area, the navigation trajectory is an Euler circuit, and the navigation distance balance is the goal.

4. The method according to claim 1, It is characterized in that The first partition set is a 2-partition set under a tree of the task network graph.

5. The method according to claim 1, It is characterized in that The calculating of the feasible solution set to obtain the network graph partitioning scheme for the drone line patrol includes: The variance of the target value of each feasible solution in the feasible solution set is calculated, and the feasible solution corresponding to the minimum variance in the variance is taken as the optimal solution output. The optimal solution is the network graph partitioning scheme for the drone line patrol.

6. A multi-UAV line patrol network graph partitioning system using the multi-UAV line patrol network graph partitioning method according to claim 1, It is characterized in that include: A generation module is used to generate a task network diagram for drone patrol; A first calculation module, configured to partition the task network graph using a first algorithm to obtain a first partition set; A second computing module, configured to screen the first partition set using a second algorithm to obtain a feasible solution set; The third calculation module is used to calculate the feasible solution set to obtain the network graph partitioning scheme for the drone line patrol.

Images