A UAV-Unmanned Vehicle Convergence Collaborative Path Planning Method and System

Abstract

The present invention belongs to the field of collaborative path planning, and provides a method and system for collaborative path planning of a drone-unmanned vehicle merging. The present invention is based on a bidirectional A* algorithm, and regards the forward search of the bidirectional A* algorithm starting from the starting point as the path search process of the drone, and the reverse search process starting from the end point as the path search process of the unmanned vehicle. The meeting position of the drone and the unmanned vehicle is determined by screening suitable meeting points, thereby realizing collaborative path planning of the drone and the unmanned vehicle. The present invention realizes cross-dimensional fusion of path search by placing the drone and the unmanned vehicle in a three-dimensional space and limiting the node expansion of the unmanned vehicle to a two-dimensional expansion. While ensuring the feasibility of the unmanned vehicle path, the optimal meeting point is selected based on the nodes obtained by the two-dimensional expansion, thereby reducing the number of candidate nodes and speeding up the path planning speed.

Description

Technical Field

[0001] The present invention belongs to the field of collaborative path planning, and specifically relates to a method and system for collaborative path planning of unmanned aerial vehicle (UAV)-unmanned vehicle merging. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] In recent years, drones have been widely used in military reconnaissance, forest rescue, agricultural plant protection, logistics and transportation, and circuit inspection due to their small size, low cost, and high efficiency. Drones can flexibly adapt to complex terrains with their wide aerial vision, but their endurance and load-bearing capacity are limited. It is difficult for a single drone to complete more complex and time-consuming tasks. Therefore, the coordinated control of drones and other robot systems, especially unmanned vehicles, has received widespread attention. Unlike drones, unmanned vehicles have obvious advantages in endurance and load-bearing capacity, and can undertake major transportation and support tasks. However, unmanned vehicles are greatly affected by terrain, and it is difficult for them to pass through complex spaces or cross some obstacles, and their perception of space is weak.

[0004] Based on the above characteristics of drones and unmanned vehicles, the collaboration between drones and unmanned vehicles can achieve complementary advantages between drones and unmanned vehicles. On the one hand, unmanned vehicles can carry drones to the work area, effectively extending the drone's hovering time; at the same time, they can also serve as temporary support platforms for drones, providing power supply, emergency landing and other functions. On the other hand, drones can provide unmanned vehicles with spatial maps and high-dimensional environmental perception, making up for the lack of environmental perception capabilities of unmanned vehicles. Reasonable collaboration can improve the mission scope and success rate of the entire system.

[0005] In the UAV-UAV collaborative system, path planning for the UAV-UAV fusion scenario is one of the basic tasks. Compared with the path planning algorithm for a single UAV, the multi-machine path planning algorithm for the UAV-UAV fusion scenario faces some new problems:

[0006] (1) The spatial dimensions considered in the path planning of drones and unmanned vehicles are different. The path planning of drones is usually carried out in three-dimensional space, while the path planning of unmanned vehicles usually only considers two-dimensional space. Therefore, the path planning algorithm for the convergence scenario of drones and unmanned vehicles needs to consider cross-dimensional issues, including cross-dimensional convergence and obstacle avoidance. (2) How to determine the convergence point. The determination of the convergence point is affected by many factors, such as the endurance of the drone and whether the unmanned vehicle is reachable, etc., which need to be considered comprehensively to obtain the desired results.

[0007] In short, path planning for the UAV and unmanned vehicle convergence scenario is not a simple extension of single-machine planning, and corresponding problems need to be solved. Summary of the invention

[0008] In order to solve the above problems, the present invention proposes a UAV-Unmanned Vehicle convergence collaborative path planning method and system. The present invention is based on a bidirectional A* algorithm, and regards the forward search of the bidirectional A* algorithm starting from the starting point as the path search process of the UAV, and the reverse search process starting from the end point as the path search process of the unmanned vehicle. The encounter position of the UAV and the unmanned vehicle is determined by screening suitable encounter points, thereby realizing collaborative path planning of the UAV and the unmanned vehicle.

[0009] According to some embodiments, the present invention adopts the following technical solutions:

[0010] A UAV-UAV fusion collaborative path planning method comprises the following steps:

[0011] Generate a grid map and select the initial positions of the UAV and the unmanned vehicle in the grid map;

[0012] Create and initialize the open list and closed list required for forward search and reverse search respectively;

[0013] At the same time, forward search and reverse search are performed alternately, wherein the forward search performs three-dimensional node expansion according to the priority order of the nodes in the forward open list, and the reverse search performs two-dimensional node expansion;

[0014] When the common nodes of the forward search and the reverse search reach the set number, the optimal meeting point is selected in the common node set based on the estimated time for the UAV along the forward search path and the UAV along the reverse search path to reach each common node;

[0015] Determine whether the selected optimal meeting point is at the boundary of the area represented by the common node set and is not close to obstacles. If so, continue searching to expand the number of common nodes until the requirement is met. If not, start from this point to retrieve the parent nodes in both directions to form a path.

[0016] As an optional implementation, the grid map contains obstacles, and the position of the drone is in the air with a certain height, the position of the unmanned vehicle is on the ground, and the height value represents the height of the unmanned vehicle.

[0017] As an optional implementation, the open list is a priority list, and the nodes in the list are sorted in ascending order according to the cost function to ensure that the node popped out from the open list each time is the node with the smallest cost;

[0018] Close the list to save only the expanded nodes.

[0019] As an optional implementation, during initialization, the starting point and the end point are respectively put into a forward open list and a reverse open list, and the node expansion direction of the reverse search is restricted so that it can only be expanded in eight directions of the plane.

[0020] As an optional implementation, during initialization, the node expansion of the reverse search is set to a two-dimensional plane expansion, and the node expansion of the forward search is set to a three-dimensional stereo expansion.

[0021] As an optional implementation, the forward search targets the initial position of the unmanned vehicle; the reverse search targets the projection of the unmanned vehicle on the plane where the unmanned vehicle is located;

[0022] The forward search and the reverse search are expanded alternately by popping out the nodes with the smallest cost functions from their respective open lists.

[0023] As a further implementation, the cost function of a node during forward search and reverse search is and They are:

[0024] ;

[0025] ;

[0026] in, Represents a node, and are the distance costs from the node to their respective starting points during forward search and backward search, and is the heuristic term, which indicates the estimated distance from the node to the respective target point during forward search and reverse search. is the weight of the heuristic term.

[0027] As a further implementation method, when the reverse search completes the search, if the overall search algorithm is not completed, the forward search continues and the bidirectional search algorithm is converted to a unidirectional search.

[0028] As a further implementation method, if the unmanned vehicle is located in an area surrounded by obstacles and cannot leave the area, but the drone can cross the obstacles and enter the area, the reverse search stops after it is unable to continue searching and waits for the result of the forward search.

[0029] As an optional implementation method, the process of selecting the optimal meeting point includes: using the node reversely searched by the unmanned vehicle as a basis for selecting the meeting point;

[0030] Assign a new cost function to all common nodes :

[0031] ;

[0032] ;

[0033] ;

[0034] in, Represents a node, Indicates the cost of the drone's hovering time, represents the time cost of UAV movement; when selecting the optimal meeting point, the movement of the UAV and the unmanned vehicle is regarded as uniform motion, and the speeds are and , and They represent the hovering time weight and the movement time weight respectively. According to this cost function, the point with the minimum cost is selected as the optimal meeting point. and They are the distance costs from the node to their respective starting points during forward search and backward search respectively.

[0035] As an optional implementation method, when the determined optimal meeting point is located on the geometric edge of the area represented by the common node set and is not adjacent to the node where the obstacle is located, the search continues and the number of common nodes is expanded until the selected meeting point is no longer on the geometric edge of the area represented by the common node set or is adjacent to the node where the obstacle is located, completing the entire search process; starting from the optimal meeting point found, the path is retrieved in reverse to the starting point and the end point, i.e., the drone and the unmanned vehicle, and the planning is completed to obtain the planned path.

[0036] A UAV-UAV fusion collaborative path planning system, comprising:

[0037] A map generation module is configured to generate a grid map and select initial positions of the UAV and the unmanned vehicle in the grid map;

[0038] A list building module is configured to respectively build and initialize an open list and a closed list required for a forward search and a reverse search;

[0039] a bidirectional search module configured to alternately perform forward search and reverse search simultaneously, wherein the forward search performs three-dimensional node expansion according to the priority order of the nodes in the forward open list, and the reverse search performs two-dimensional node expansion;

[0040] The optimal meeting point selection module is configured to select the optimal meeting point in the common node set according to the estimated time for the UAV to reach each common node along the forward search path and the UAV to reach each common node along the reverse search path when the common nodes of the forward search and the reverse search reach a set number;

[0041] The path planning module is configured to determine whether the selected optimal meeting point is at the boundary of the area represented by the common node set and is not close to obstacles. If so, continue searching to expand the number of common nodes until the requirements are met. If not, start from this point to retrieve the parent nodes in two directions to form a path.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] The present invention is based on the bidirectional A* algorithm, and regards the forward search of the bidirectional A* algorithm starting from the starting point as the path search process of the UAV, and the reverse search process starting from the end point as the path search process of the unmanned vehicle. Through cross-dimensional search fusion, finding the optimal meeting point with the goal of minimizing the time consumption of the UAV, and providing an optimal meeting point optimization method, the optimal meeting point obtained by the improved algorithm is used as the common target point of the UAV and the unmanned vehicle, realizing cross-dimensional collaborative path planning in the convergence scenario of the UAV and the unmanned vehicle.

[0044] The present invention comprehensively considers the problems faced by drones and unmanned vehicles in collaborative path planning in convergence scenarios, improves planning efficiency and the degree of coordination between drones and unmanned vehicles; by placing drones and unmanned vehicles in a three-dimensional space and limiting the node expansion of the unmanned vehicle to a two-dimensional expansion, cross-dimensional fusion of path search is achieved, and while ensuring the feasibility of the unmanned vehicle path, the optimal meeting point is selected based on the nodes obtained by the two-dimensional expansion, which reduces the number of alternative nodes and speeds up path planning.

[0045] The present invention binds the end flags of the reverse search and the forward search. The entire search process is completed only after the slower forward search completes the requirement, so that the algorithm can be applied in an environment where an unmanned vehicle is surrounded by obstacles, thereby expanding the scope of application.

[0046] The present invention establishes a new cost function for selecting the optimal meeting point, introduces the time-consuming cost of drone hovering and the time-consuming cost of movement, and reduces the time consumption of the fusion of drone and unmanned vehicle; provides an optimal meeting point optimization method, by detecting the edges of the common node area, the optimal solution of the area is obtained in an iterative manner.

[0047] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0049] Figure 1A schematic diagram of a process flow of UAV-UAV merging collaborative path planning based on a bidirectional A* algorithm according to an embodiment;

[0050] Figure 2 An algorithm flow chart of an improved path planning algorithm according to an embodiment;

[0051] Figure 3 A result diagram showing an obstacle environment obtained in one embodiment;

[0052] Figure 4 A result diagram of an embodiment in which obstacles are not visualized;

[0053] Figure 5 This is a diagram of path planning results obtained in an embodiment when the unmanned vehicle is in a trapped state. DETAILED DESCRIPTION

[0054] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0055] It should be noted that the following detailed descriptions are all illustrative and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0056] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0057] In the absence of conflict, the embodiments in this application and the features in the embodiments may be combined with each other.

[0058] Embodiment 1

[0059] A UAV-UAV fusion collaborative path planning method based on a bidirectional A* algorithm, such as Figure 1 As shown, the following steps are included:

[0060] Step 1: Create a grid map and give the initial positions of the unmanned vehicle and the drone. Figure 3 As shown, a task environment with dense obstacles is simulated.

[0061] This embodiment is aimed at path planning in the scenario where a drone and an unmanned vehicle meet, the drone should be in the air with a certain height, and the unmanned vehicle should be on the ground, and the height value it has represents the height of the unmanned vehicle. Figure 3The red balls at both ends of the path in the figure represent the drone and the unmanned vehicle respectively.

[0062] Step 2: Initialize the lists and values ​​required by the algorithm, and create open lists and closed lists for forward search and reverse search respectively to store nodes.

[0063] The open lists are all priority lists. In this embodiment, they are sorted from small to large according to the cost function to ensure that the node popped out from the open list each time is the node with the smallest cost.

[0064] The closed list is a normal list to ensure that it will not be expanded repeatedly. There is no need for priority sorting, and only query operations are performed.

[0065] During the initialization process, the starting point and the end point need to be put into the forward open list and the reverse open list respectively. In addition, because the unmanned vehicle only considers the movement on the plane, in order to ensure that the path obtained by the reverse search representing the path planning of the unmanned vehicle is feasible, the node expansion direction of the reverse search is limited so that it can only be expanded in eight directions of the plane.

[0066] During the initialization process, the number of common nodes is limited. Only when the number requirement is met will the optimal meeting point selection process be entered, unless the forward search has reached the target point before the number requirement is met. This situation usually occurs when the initial position of the drone and the unmanned vehicle is very close. At this time, the optimal meeting point is selected from the common nodes that have been obtained.

[0067] Figure 3-5 In the experiment shown, the blue translucent blocks represent the common nodes detected during the planning process. The initial number of common nodes is 20, and the step size needs to be increased to 20.

[0068] Step 3: If Figure 2 As shown in the figure, the forward search targets the initial position of the unmanned vehicle; the reverse search targets the projection of the drone on the plane where the unmanned vehicle is located. The forward search will perform three-dimensional node expansion according to the priority order of the nodes in the forward open list. The node expansion will ignore the points that have been expanded and the points occupied by obstacles. The expanded points will be put into the forward closed list. The expansion method of the reverse search is the same as that of the forward search, except that the reverse search performs two-dimensional node expansion.

[0069] The forward search and reverse search are alternately expanded by popping up the nodes with the smallest cost function in their respective open lists. The cost function of the node in the forward search and reverse search is and They are:

[0070] ;

[0071] ;

[0072] in, Represents a node, and are the distance costs from the node to their respective starting points during forward search and backward search, and It is a heuristic term, which indicates the estimated distance from the node to the respective target point during forward search and reverse search, generally expressed as Euclidean distance. is the weight of the heuristic term.

[0073] In addition, because the forward search is a three-dimensional search, its search speed is much lower than the two-dimensional reverse search. When the open list of the forward search is empty and has not been completed, it means that the drone cannot find a feasible path to merge with the unmanned vehicle, which is considered a path planning failure.

[0074] In this embodiment, the reverse search stops after it completes the search, but the entire planning continues. At this time, the two-way search algorithm is converted to a one-way search. In addition, if the unmanned vehicle is located in an area surrounded by obstacles, the unmanned vehicle cannot leave the area, but the drone can cross the obstacles and enter the area. In this case, the reverse search cannot reach the set target point. After traversing all feasible nodes, the reverse search will not feedback the planning failure, but stop the search and wait for the results of the forward search. This effectively expands the applicable scenarios of the algorithm.

[0075] In this process, no matter what the result of the reverse search is, the forward search process will not be interrupted. Through this design, the algorithm is suitable for Figure 5 The figure shows the path planning of the unmanned vehicle in a besieged environment. In this scenario, the two-dimensional planning of the unmanned vehicle cannot reach the destination. After traversing the entire enclosed space, that is, after the reverse open list is empty, the entire planning does not end, but waits for the forward three-dimensional search to cross the obstacle and complete the planning.

[0076] During the search process in each direction, in order to avoid repeated expansion of nodes, the respective closed lists in each direction are searched. If the node exists in the closed list in the search direction, it means that the node has been expanded in this search direction, so it is skipped.

[0077] This process does not consider whether the node exists in the closed list in other directions, so that the common node can be retrieved according to the intersection of the closed lists in different directions. Then, by excluding collision points and out-of-bounds points, feasibility is ensured, and the remaining points that need to be expanded are retained. Each node has two costs, one for the forward direction and the other for the reverse direction.

[0078] When expanded by a forward or reverse search, it will be used according to the corresponding direction. or The cost in the search direction is updated and imported into the respective open lists, thereby realizing forward search and reverse search.

[0079] Step 4: As mentioned in step 3, since only the closed lists of each direction are considered during the expansion process, the closed lists of different directions are independent of each other. Therefore, the common nodes can be detected by simply taking the intersection. When the number of common nodes detected is greater than or equal to the set value, Figure 3-5 Taking the experiment shown as an example, when the number of nodes exceeds 20, the optimal meeting point selection process will be triggered. This process will create a new priority queue to store all the current common nodes, and calculate a new cost function for each node. , sorting the drones with the optimal time consumption as the goal, and then finding the optimal meeting point.

[0080] The optimal point selection method assigns a new cost function to all common nodes :

[0081] ;

[0082] ;

[0083] ;

[0084] in, Indicates the cost of the drone's hovering time, The cost of the UAV’s movement time. Compared with UAVs, the unmanned vehicle has a significant advantage in endurance, so the selection of the optimal point mainly considers the relevant time of the UAV, and does not consider the time of the unmanned vehicle. In order to simplify the calculation, the movement of the UAV and the unmanned vehicle is regarded as uniform motion, and the speed is and . and Respectively represent the hovering time weight and the movement time weight. In this embodiment, in the simulation experiment, the weight value and The value is 1.1.

[0085] Finally, the point with the lowest cost is selected as the optimal meeting point according to this cost function.

[0086] If the forward search is completed during the planning process, but the number of common nodes has not yet reached the required number, the common nodes currently found will be selected in the same way. This usually occurs when the initial distance between the drone and the unmanned vehicle is very short. At this time, the number of nodes is very small, but the proportion of common nodes in all expanded nodes is high, and there is no need to further judge the meeting point, such as Figure 2 shown.

[0087] Step 5: In the first case described in step 4, the optimal meeting point may be selected at the geometric edge of the area represented by the common node set and is not adjacent to the node where the obstacle is located, that is, Figure 3 or Figure 5 In this case, there may be a better point in this direction, but this point is not among the candidate points, making the selected point the current optimal point. Therefore, it is necessary to expand the range of candidate points, i.e., common nodes. Figure 2 As shown in the figure, by expanding the number of common nodes, the optimal meeting point can jump out of the local optimum and obtain a better result.

[0088] Finally, path planning is achieved by starting from the optimal meeting point and continuously finding the parent nodes of the nodes obtained in the forward and reverse searches to find the complete path. If the forward search traverses the entire accessible space of the drone, that is, the forward open list is empty, and the end condition is not met, it is considered that the situation cannot be planned and the planning fails.

[0089] Figure 3 This is the experimental result showing obstacles. The red ball with height on the left side of the picture represents the drone, the one on the right side on the ground represents the unmanned vehicle, the blue ball with a yellow dot in the middle represents the optimal meeting point, the path is represented by a red line, and the translucent blue squares represent common nodes. Figure 4 yes Figure 3 It can be seen that the algorithm proposed in the present invention realizes the coordinated path planning of UAV and vehicle fusion in complex environments and facing different situations, and the effect is good.

[0090] Embodiment 2

[0091] A UAV-UAV fusion collaborative path planning system, comprising:

[0092] A map generation module is configured to generate a grid map and select initial positions of the UAV and the unmanned vehicle in the grid map;

[0093] The execution process of this module can be specifically referred to the process of step 1 of the first embodiment, which will not be repeated here.

[0094] A list building module is configured to respectively build and initialize an open list and a closed list required for a forward search and a reverse search;

[0095] The execution process of this module can be specifically referred to the process of step 2 of the first embodiment, which will not be repeated here.

[0096] a bidirectional search module configured to alternately perform forward search and reverse search simultaneously, wherein the forward search performs three-dimensional node expansion according to the priority order of the nodes in the forward open list, and the reverse search performs two-dimensional node expansion;

[0097] The execution process of this module can be specifically referred to the process of step 3 of the first embodiment, which will not be repeated here.

[0098] The optimal meeting point selection module is configured to select the optimal meeting point in the common node set according to the estimated time for the UAV to reach each common node along the forward search path and the UAV to reach each common node along the reverse search path when the common nodes of the forward search and the reverse search reach a set number;

[0099] The execution process of this module can be specifically referred to the process of step 4 of the first embodiment, which will not be repeated here.

[0100] The path planning module is configured to determine whether the selected optimal meeting point is at the boundary of the area represented by the common node set and is not close to obstacles. If so, continue searching to expand the number of common nodes until the requirements are met. If not, start from this point to retrieve the parent nodes in two directions to form a path.

[0101] The execution process of this module can be specifically referred to the process of step 5 of the first embodiment, which will not be repeated here.

[0102] It can be understood that the above-mentioned units / modules can be separately or completely combined into one or several other units / modules, or one (some) of the units can be further divided into multiple functionally smaller units, which can achieve the same operation without affecting the realization of the technical effects of the embodiments of the present application.

[0103] The above modules of the system are divided based on logical functions. In practical applications, the function of one module can also be implemented by multiple modules, or the functions of multiple modules can be implemented by one module.

[0104] Similarly, in other embodiments of the present application, the system may also include other units / modules. In actual applications, these functions may also be implemented with the assistance of other units, and may be implemented by collaboration of multiple units.

[0105] According to another embodiment of the present application, the system described in this embodiment can be constructed and the method of Embodiment 1 can be implemented by running a computer program (including program code) capable of executing the steps involved in the corresponding method described in Embodiment 1 on a general-purpose computing device such as a computer, which includes processing elements and storage elements such as a central processing unit (CPU), random access memory (RAM), and read-only memory (ROM). The computer program can be recorded on, for example, a computer-readable recording medium, and loaded into the above-mentioned computing device through the computer-readable recording medium and run therein.

[0106] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made by those skilled in the art within the spirit and principle of the present invention without creative labor shall be included in the protection scope of the present invention.

Claims

1. A UAV-UAV fusion collaborative path planning method, characterized in that: The following steps are involved: Generate a grid map and select the initial positions of the UAV and the unmanned vehicle in the grid map; Create and initialize the open list and closed list required for forward search and reverse search respectively; At the same time, forward search and reverse search are performed alternately, wherein the forward search performs three-dimensional node expansion according to the priority order of the nodes in the forward open list, and the reverse search performs two-dimensional node expansion; When the number of common nodes in the forward search and reverse search reaches the set number, the optimal meeting point is selected in the common node set based on the estimated time for the UAV to reach each common node along the forward search path and the unmanned vehicle along the reverse search path; when selecting the optimal meeting point, the node searched by the unmanned vehicle in reverse is used as the basis for selecting the meeting point. Specifically, a new cost function is assigned to all common nodes. ,Right now: ； ； ； in, Represents a node, Indicates the cost of the drone's hovering time, represents the time cost of the UAV movement; when selecting the optimal meeting point, the movement of the UAV and the unmanned vehicle is regarded as uniform motion, and the speeds are and , and They represent the hovering time weight and the movement time weight respectively. According to this cost function, the point with the minimum cost is selected as the optimal meeting point. and are the distance costs from the node to the respective starting points during forward search and backward search respectively; Determine whether the selected optimal meeting point is at the boundary of the area represented by the common node set and is not close to obstacles. If so, continue searching to expand the number of common nodes until the requirement is met. If not, start from this point to retrieve the parent nodes in both directions to form a path.

2. A UAV-UAV fusion collaborative path planning method as claimed in claim 1, characterized in that: The grid map contains obstacles, and the position of the drone is in the air with a certain height, and the position of the unmanned vehicle is on the ground, and the height value represents the height of the unmanned vehicle.

3. The method for planning a coordinated path for a UAV-UAV fusion according to claim 1, characterized in that: The open list is a priority list. The nodes in the list are sorted from small to large according to the cost function to ensure that the node popped out from the open list each time is the node with the lowest cost; Close the list to save only the expanded nodes.

4. The method for planning a coordinated path for a UAV-AV fusion according to claim 1, characterized in that: During initialization, the starting point and the end point are placed in the forward open list and the reverse open list respectively, and the node expansion direction of the reverse search is restricted so that it can only be expanded in the eight directions of the plane; Or, during initialization, the node expansion of the reverse search is set to a two-dimensional plane expansion, and the node expansion of the forward search is set to a three-dimensional stereo expansion.

5. The method for planning a coordinated path for a UAV-UAV fusion according to claim 1, characterized in that: The forward search targets the initial position of the unmanned vehicle; the reverse search targets the projection of the drone on the plane where the unmanned vehicle is located; The forward search and the reverse search are expanded alternately by popping out the nodes with the smallest cost functions from their respective open lists.

6. A UAV-UAV fusion collaborative path planning method as claimed in claim 5, characterized in that: Cost function of a node during forward and reverse search and They are: ； ； in, Represents a node, and are the distance costs from the node to their respective starting points during forward search and reverse search, and is the heuristic term, which indicates the estimated distance from the node to the respective target point during forward search and reverse search. is the weight of the heuristic term.

7. A UAV-UAV fusion collaborative path planning method as claimed in claim 6, characterized in that: When the reverse search completes its search, if the overall search is not completed, the forward search continues and the bidirectional search algorithm is converted to a one-way search; Or, if the UAV is in an area surrounded by obstacles and the UAV cannot leave the area, but the UAV can cross the obstacles and enter the area, then stop the reverse search and wait for the result of the forward search.

8. The method for planning a coordinated path for a UAV-AV fusion according to claim 1, characterized in that: The process of selecting the optimal meeting point includes: using the nodes searched by the unmanned vehicle in reverse as the basis for selecting the meeting point; Assign a new cost function to all common nodes : ； ； ； in, Represents a node, Indicates the cost of the drone's hovering time, represents the time cost of UAV movement; when selecting the optimal meeting point, the movement of the UAV and the unmanned vehicle is regarded as uniform motion, and the speeds are and , and They represent the hovering time weight and the movement time weight respectively. According to this cost function, the point with the minimum cost is selected as the optimal meeting point. and They are the distance costs from the node to their respective starting points during forward search and backward search respectively.

9. The method for planning a coordinated path for a UAV-UAV fusion according to claim 1, characterized in that: When the determined optimal meeting point is located on the geometric edge represented by the common node set and is not adjacent to the node where the obstacle is located, the search continues and the number of common nodes is expanded until the selected meeting point is not on the geometric edge of the common node set or is adjacent to the node where the obstacle is located, completing the entire search process; starting from the found optimal meeting point, the path is retrieved in reverse to the starting point and the end point, i.e., the drone and the unmanned vehicle, and the planning is completed to obtain the planned path.

10. A UAV-UAV fusion collaborative path planning system, characterized by: include: A map generation module is configured to generate a grid map and select initial positions of the UAV and the unmanned vehicle in the grid map; A list building module is configured to respectively build and initialize an open list and a closed list required for a forward search and a reverse search; a bidirectional search module configured to alternately perform forward search and reverse search simultaneously, wherein the forward search performs three-dimensional node expansion according to the priority order of the nodes in the forward open list, and the reverse search performs two-dimensional node expansion; The optimal meeting point selection module is configured to select the optimal meeting point in the common node set according to the estimated time for the UAV to reach each common node along the forward search path and the unmanned vehicle along the reverse search path when the common nodes of the forward search and reverse search reach the set number; when selecting the optimal meeting point, the node searched by the unmanned vehicle in the reverse direction is used as the basis for selecting the meeting point. Specifically, a new cost function is assigned to all common nodes. ,Right now: ； ； ； in, Represents a node, Indicates the cost of the drone's hovering time, represents the time cost of the UAV movement; when selecting the optimal meeting point, the movement of the UAV and the unmanned vehicle is regarded as uniform motion, and the speeds are and , and They represent the hovering time weight and the movement time weight respectively. According to this cost function, the point with the minimum cost is selected as the optimal meeting point. and are the distance costs from the node to the respective starting points during forward search and backward search respectively; The path planning module is configured to determine whether the selected optimal meeting point is at the boundary of the area represented by the common node set and is not close to obstacles. If so, continue searching to expand the number of common nodes until the requirements are met. If not, start from this point to retrieve the parent nodes in two directions to form a path.

Images