Multi-agent path planning method, device and equipment

Abstract

The invention discloses a multi-agent path planning method, device and equipment, and the method comprises the steps: dividing a map according to first grids of a preset size, carrying out the global planning of all agents on the map, and generating a global path, the global path being a road section passing through a plurality of first grids from a starting point to a target point; on the basis of each path section of the global path, local planning is conducted on all agents, a local path is generated, and each path section is a road section formed by the center points of every two adjacent first grids; real-time data are obtained by monitoring changes of dynamic obstacles and dynamic targets in a map in real time, and the local path and / or the local path are / is updated and planned according to the real-time data to obtain an updated path. Potential collision risks can be quickly found, obstacle avoidance measures can be taken in time, and the safety of a multi-agent system and the smooth completion rate of tasks are improved.

Description

Technical Field

[0001] The present invention relates to the field of path planning technology, and in particular to a multi-agent path planning method, device and equipment. Background Art

[0002] In Multi-Agent Systems (MAS), Multi-Agent Path Finding (MAPF) is a crucial research field, especially in the problem of missile path planning, where its complexity is particularly significant. The core task of missile path planning is to accurately plan an optimal path from the starting point to the target point for the missile in a complex and changing combat environment. This process not only requires high-precision modeling of the environment, but also requires full consideration of the kinematic characteristics of the missile, such as constraints such as speed, acceleration, and steering ability. At the same time, missile path planning must also be highly real-time to cope with the ever-changing battlefield and ensure the coordinated operation of multiple missiles to avoid path conflicts and achieve precision strikes.

[0003] Existing multi-agent path planning methods, such as rule-based methods, optimization-based methods, and search-based methods, all provide solutions to a certain extent. Among them, the conflict-based search (CBS) algorithm has been widely used in the field of multi-agent path planning due to its high efficiency and optimization. The CBS algorithm effectively reduces the computational complexity of path planning by first independently planning a path for each agent and then locally adjusting the conflict resolution strategy. However, with the increase in the number of agents and the increasing complexity of the battlefield environment, the CBS algorithm has also exposed some obvious limitations. For example, Scalability issue: As the number of missiles increases, the computational complexity of the CBS algorithm rises sharply, resulting in a significant decrease in the algorithm's execution efficiency and performance in large-scale missile path planning tasks, limiting its application in complex combat environments.

[0004] Insufficient real-time performance: In a highly dynamic battlefield environment, missile path planning needs to respond quickly to environmental changes, but the CBS algorithm is not efficient enough in path recalculation and cannot meet the real-time requirements of missile path planning.

[0005] Optimization limitations: The traditional CBS algorithm does not perform well in key optimization indicators such as path length and time, which affects the overall effectiveness and strike accuracy of missile path planning. Summary of the invention

[0006] In view of this, the purpose of the present invention is to propose a multi-agent path planning method, device and equipment, aiming to solve how to improve the scalability, real-time and optimization performance of the multi-agent path planning algorithm in missile path planning in order to cope with complex and changeable combat environments.

[0007] To achieve the above object, the present invention provides a multi-agent path planning method, the method comprising: Divide the map into first grids of a preset size, and perform global planning for all agents on the map to generate a global path, wherein the global path is a section from a starting point to a target point that passes through multiple first grids; Performing local planning for all agents based on each path segment of the global path to generate a local path, wherein each path segment is a path segment formed by the center points of two adjacent first grids; By real-time monitoring of changes in dynamic obstacles and dynamic targets in the map, real-time data is obtained, and the local path and / or the local path are updated and planned according to the real-time data to obtain an updated path.

[0008] Preferably, the planning process of the global planning / the local planning includes: Use the Hybird A-star algorithm to perform path planning and obtain the initial path; Performing conflict detection on the initial path, and performing conflict resolution on the detected path conflict through a pre-built conflict tree to obtain a conflict-free path; The global path / the local path is obtained by recursively calling the CBS algorithm to perform repeated conflict detection and conflict processing on the conflict-free path.

[0009] Preferably, the conflict tree construction process includes: Initialize the root node of the conflict tree, where the root node includes the initial paths of all agents; When a path conflict is detected, the path conflict is added to a conflict node in a conflict tree, wherein each conflict node represents a planning problem of a path segment.

[0010] Preferably, the step of performing conflict handling on the detected path conflict by using a pre-built conflict tree comprises: Creating two child nodes based on the conflict node in the conflict tree for the detected path conflict to obtain a first child node and a second child node, wherein each child node indicates that an agent in the path conflict is prohibited from appearing at the conflict position; The conflicting position is avoided by re-planning a path for the first agent on the first child node, and the conflicting position is avoided by re-planning a path for the second agent on the second child node.

[0011] Preferably, dynamic constraints and resource constraints are introduced into the Hybrid A-star algorithm to set the cost function.

[0012] Preferably, the cost function is defined as , where g(x) represents the actual cost from the starting point to the current path point, and h(x) represents the heuristic estimate from the current path point to the target path point. , where ɑ, β, and γ represent the weight of the path length cost PATH_L, the weight of the corner cost TURN_CT, and the weight of the maximum path cost LIMIT_L respectively; , Among them, x i ,y i 、z i Indicates the location information of the current path point, x i+1 ,y i+1 、z i+1 Indicates the location information of the next path point; ,in, , , where represents the corner calculation cost, k represents the weight coefficient, i Indicates the heading angle of the current path point. i+1 Indicates the heading angle of the next path point, R min represents the minimum turning radius, v represents the speed, represents the maximum angular velocity; , where L max Indicates the maximum path length; , x i ,y i 、z i Indicates the location information of the current path point, x goal ,y goal 、z goal Indicates the location information of the target waypoint.

[0013] Preferably, the updating and planning of the local path and / or the local path according to the real-time data includes: Determine the location of the path conflict according to the real-time data, and if the path conflict is located on the global path, update the planning of the global path and the local path; If the path conflict is on the local path, the local path is updated and planned.

[0014] To achieve the above object, the present invention further provides a multi-agent path planning device, the device comprising: A global planning unit, used to divide the map into first grids of a preset size, and perform global planning for all agents on the map to generate a global path, wherein the global path is a section from a starting point to a target point passing through a plurality of the first grids; A local planning unit, configured to perform local planning for all agents based on each path segment of the global path to generate a local path, wherein each path segment is a path segment formed by the center points of two adjacent first grids; The path updating unit is used to obtain real-time data by monitoring the changes of dynamic obstacles and dynamic targets in the map in real time, and to update and plan the local path or / and the local path according to the real-time data to obtain an updated path.

[0015] In order to achieve the above objectives, the present invention also proposes a multi-agent path planning device, including a processor, a memory, and a computer program stored in the memory, wherein the computer program is executed by the processor to implement the steps of a multi-agent path planning method as described in the above embodiment.

[0016] In order to achieve the above objectives, the present invention also proposes a computer-readable storage medium, on which a computer program is stored, and the computer program is executed by a processor to implement the steps of a multi-agent path planning method as described in the above embodiment.

[0017] In order to achieve the above objectives, the present invention also proposes a computer program product, including a computer program / instructions, which, when executed by a processor, implement the steps of a multi-agent path planning method as described in the above embodiment.

[0018] Beneficial effects: The above scheme divides the map into a first grid of a preset size, performs global planning for all intelligent agents, generates a global path, performs local planning based on each path segment of the global path, and generates a local path. Global planning ensures the rationality of the overall path, while local planning improves the accuracy of the path. The combination of global and local planning helps to improve the flexibility and adaptability of path planning in complex and changeable combat environments; furthermore, through real-time monitoring and updating, it can cope with changes in a dynamic environment and improve the robustness and real-time performance of path planning.

[0019] The above scheme uses the Hybrid A-star algorithm for path planning in global and local planning, then performs conflict detection and conflict resolution, and finally obtains a conflict-free path by recursively calling the CBS algorithm. The pre-built conflict tree and CBS algorithm effectively solve the conflict problem in multi-agent path planning. The structured conflict tree provided is used to manage and resolve path conflicts, making conflict resolution clearer and more efficient. The combination of the Hybrid A-star algorithm and the CBS algorithm can quickly generate an initial path, effectively resolve path conflicts, and improve the efficiency and optimization of path planning.

[0020] The above scheme introduces dynamic constraints and resource constraints in the Hybrid A-star algorithm to set the cost function, making the path planning more in line with the actual application scenario and improving the efficiency and safety of path planning. By defining a cost function that comprehensively considers the path length, corner cost and maximum path cost, it is possible to more comprehensively evaluate the pros and cons of the path, optimize the path selection and improve the performance of path planning.

[0021] The above scheme, by constructing a conflict tree to handle path conflicts, can systematically manage and resolve conflicts between agents, improving the efficiency of conflict handling; the use of conflict trees makes the conflict handling process more structured, which helps to quickly locate and resolve conflicts. Updating and planning local or global paths based on real-time data can quickly respond to environmental changes and improve the adaptability of path planning to dynamic environments. However, selectively updating the global path or local path based on the conflict location improves the update efficiency and path continuity. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0023] Figure 1 A flowchart of a multi-agent path planning method provided by one embodiment of the present invention.

[0024] Figure 2 A schematic diagram of the structure of a multi-agent path planning device provided by one embodiment of the present invention.

[0025] The realization of the purpose, functional features and advantages of the invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the invention claimed for protection, but merely represents the selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0027] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features.

[0028] The content of the present invention is described in detail below with reference to embodiments.

[0029] Reference Figure 1 Shown is a flow chart of a multi-agent path planning method provided by an embodiment of the present invention.

[0030] In this embodiment, the method includes: S11, dividing the map into first grids of a preset size, and performing global planning for all agents on the map to generate a global path, wherein the global path is a road section from a starting point to a target point passing through multiple first grids.

[0031] In this embodiment, the method is applied to path planning of missile multi-agents. Specifically, by creating a global plan, the global plan is performed on a larger grid to determine the approximate path from the starting point to the target point. It includes dividing the map into larger grids (the size of the larger grid is relatively large, such as 10KM, 20KM) to create a global map, performing global path planning for all missile agents based on the global map, and generating a rough path from the starting point to the end point; further, after finding a conflict-free path in the future, the path can be further optimized to reduce the total path length or other cost function values. According to the grid method (grid method) used to model the entire map, such as in units of 10KM, determine the large grid from the starting point A to the end point B, and the center point of each grid is recorded as P1, P2, P3...Pn to achieve global planning. The global planning does not need to be very detailed, and some details can be ignored. It is only necessary to ensure the general direction of the path and avoid large obstacles.

[0032] S12, performing local planning for all agents based on each path segment of the global path to generate a local path, wherein each path segment is a segment formed by center points of two adjacent first grids.

[0033] In this embodiment, by creating a local plan, the local plan is performed on a smaller grid to refine each path segment of the global path, handle local obstacles and details, and ensure the feasibility and smoothness of the path. It includes creating a local map according to each path segment of the global path, performing local path planning for all missile agents based on the local map, and generating refined path segments; further, after finding a conflict-free path, the path can be further optimized to reduce the total path length or other cost function values. For example, the path segment from P1 (Pn) to P2 (Pn+1) is further locally planned, that is, the two larger grids are divided into smaller grids using the grid method, and path planning is performed based on the smaller grids to generate path points P11, P12, P13... and so on.

[0034] Furthermore, in step S11 or S12, the planning process of the global planning / the local planning includes: Use the Hybird A-star algorithm to perform path planning and obtain the initial path; Performing conflict detection on the initial path, and performing conflict resolution on the detected path conflict through a pre-built conflict tree to obtain a conflict-free path; The global path / the local path is obtained by recursively calling the CBS algorithm to perform repeated conflict detection and conflict processing on the conflict-free path.

[0035] Furthermore, the conflict tree construction process includes: Initialize the root node of the conflict tree, where the root node includes the initial paths of all agents; When a path conflict is detected, the path conflict is added to a conflict node in a conflict tree, wherein each conflict node represents a planning problem of a path segment.

[0036] Furthermore, the detected path conflict is processed through a pre-built conflict tree, including: Creating two child nodes based on the conflict node in the conflict tree for the detected path conflict to obtain a first child node and a second child node, wherein each child node indicates that an agent in the path conflict is prohibited from appearing at the conflict position; The conflicting position is avoided by re-planning a path for the first agent on the first child node, and the conflicting position is avoided by re-planning a path for the second agent on the second child node.

[0037] In this embodiment, the implementation process of generating a rough path (global path) by global planning and generating a refined path (local path) by local planning is basically the same, specifically including: (1) Use the Hybrid A-star algorithm to perform path planning for all missile agents and generate the initial path from the start point to the end point; (2) Detect conflicts on the initial path, resolve the detected path conflicts through the conflict tree, and obtain a conflict-free path; path conflicts usually refer to two agents arriving at the same location at the same time, or two agents crossing adjacent locations in the same time period; (3) Constructing a conflict tree, including: creating a root node, initializing the root node of the conflict tree, where the root node contains the initial paths of all agents; creating conflict nodes: when a conflict is detected, adding the conflict to the conflict node in the conflict tree, each conflict node represents a planning problem for a path segment; (4) Conflict handling, including: for each detected conflict, creating two child nodes based on the conflict node in the conflict tree, each child node prohibiting one of the agents in the conflict from appearing at the conflict position; replanning the path for the first agent through child node 1 to avoid the conflict position, and replanning the path for the second agent through child node 2 to avoid the conflict position; (5) Recursive search: Recursively call the CBS algorithm and recursively apply the CBS algorithm to each child node to perform repeated conflict detection and conflict resolution until all paths are conflict-free or it is determined that the problem has no solution.

[0038] S13, obtaining real-time data by real-time monitoring the changes of dynamic obstacles and dynamic targets in the map, and updating and planning the local path and / or the local path according to the real-time data to obtain an updated path.

[0039] In this embodiment, real-time data is obtained by monitoring environmental changes in real time, including detecting changes in dynamic obstacles and dynamic targets in the map; and the local path and / or local path are replanned according to the real-time data to ensure that the updated path meets the current environment and mission requirements. For example, after the initial global path and local path planning, the missile flies along the specified path. During the flight, it encounters obstacles such as drones, resulting in conflicts with the original path. At this time, it is necessary to make a judgment on the conflicting section. If the conflict point is on the global path, the global path and local path need to be replanned; if the conflict point is on the local path, the local path is replanned.

[0040] Furthermore, dynamic constraints and resource constraints are introduced into the Hybrid A-star algorithm to set the cost function; wherein, The cost function is defined as , where g(x) represents the actual cost from the starting point to the current path point, and h(x) represents the heuristic estimate from the current path point to the target path point. , where ɑ, β, and γ represent the weight of the path length cost PATH_L, the weight of the corner cost TURN_CT, and the weight of the maximum path cost LIMIT_L respectively; , Among them, x i ,y i 、z i Indicates the location information of the current path point, x i+1 ,y i+1 、z i+1 Indicates the location information of the next path point; ,in, , , where represents the corner calculation cost, k represents the weight coefficient, i Indicates the heading angle of the current path point. i+1 Indicates the heading angle of the next path point, R min represents the minimum turning radius, v represents the speed, represents the maximum angular velocity; , where L max represents the maximum path length cost; , x i ,y i 、z i Indicates the location information of the current path point, x goal ,y goal 、z goal Indicates the location information of the target waypoint (end point).

[0041] In this embodiment, the Hybrid A-star algorithm combines discrete grid search and continuous state space search, and is applicable to a multi-missile system with non-holonomic motion constraints. The Hybrid A-star algorithm in this embodiment sets the cost function by comprehensively considering dynamic constraints and resource constraints. The information of multi-agents (such as missiles, aircraft, and other multi-agents involving corner costs) includes: State space representation: , where x, y and z are position information, is the heading angle; minimum turning radius: , where v is the velocity, is the maximum angular velocity.

[0042] The hybrid A algorithm uses the standard A cost function by combining path length and heuristic estimation. That is, the cost function f(x) is defined as: , Among them, g(x) is the actual cost from the starting point to the current node (current path point), and h(x) is the heuristic estimate from the current node to the target node (target path point).

[0043] The actual cost includes path length cost and corner cost: , Among them, ɑ, β, and γ represent the weight of the path length cost, the weight of the corner cost, and the weight of the maximum path cost, respectively.

[0044] The path length cost is: , Among them, x i ,y i 、z i Indicates the location information of the current path point, x i+1 ,y i+1 、z i+1 Indicates the location information of the next waypoint.

[0045] The corner cost is: , in, is the turning calculation cost, considering the missile agent’s minimum turning radius and turning angle.

[0046] The corner calculation cost is: , where k is the weight coefficient.

[0047] The maximum path cost is: , Among them, L max is the maximum path length, and the weight λ is taken to be large, which is used to significantly increase the cost of exceeding the path length constraint.

[0048] Heuristic estimation, generally calculated using Euclidean distance: , x i ,y i 、z iThey are the horizontal axis, vertical axis and vertical axis, and can also refer to longitude, latitude and altitude.

[0049] In this embodiment, by adopting efficient conflict detection and dynamic planning, potential collision risks can be quickly discovered and obstacle avoidance measures can be taken in time, which improves the safety of the multi-agent system and the smooth completion rate of the task; it can complete the task more efficiently and with higher quality, reduce the error rate and waste of resources, and improve the efficiency and quality of task completion. By using advanced communication protocols and collaborative strategies, real-time information sharing and coordinated decision-making between agents are realized, and the collaborative combat capability between agents is enhanced (the CBS algorithm indirectly realizes the collaboration between agents by complying with global constraints in multi-agent path planning to avoid conflicts), especially in multi-agent systems, which significantly improves the efficiency and flexibility of the overall system. The system has adaptive learning and optimization capabilities, and can continuously adjust and optimize strategies according to the environment and task requirements, so that the system performs better in a dynamically changing environment, improves the performance of the system, and enhances its robustness and reliability; because the system has strong adaptive learning and optimization capabilities, it can flexibly respond to various situations in a complex and changing environment, has a wide range of applications, and has high application value.

[0050] Reference Figure 2 Shown is a schematic structural diagram of a multi-agent path planning device provided by an embodiment of the present invention.

[0051] In this embodiment, the device 20 includes: A global planning unit 21, configured to divide the map into first grids of a preset size, and perform global planning for all agents on the map to generate a global path, wherein the global path is a section from a starting point to a target point passing through a plurality of the first grids; A local planning unit 22, configured to perform local planning for all agents based on each path segment of the global path to generate a local path, wherein each path segment is a path segment formed by the center points of two adjacent first grids; The path updating unit 23 is used to obtain real-time data by monitoring the changes of dynamic obstacles and dynamic targets in the map in real time, and update the local path and / or the local path according to the real-time data to obtain an updated path.

[0052] Each unit module of the device 20 can respectively execute the corresponding steps in the above method embodiment, so each unit module will not be described in detail here. Please refer to the description of the above corresponding steps for details.

[0053] The embodiment of the present invention further provides a multi-agent path planning device, which includes the multi-agent path planning device as described above, wherein the multi-agent path planning device can adopt Figure 2 The structure of the embodiment can be executed accordingly. Figure 1 The technical solution of the method embodiment shown has similar implementation principles and technical effects. For details, please refer to the relevant records in the above embodiments, which will not be repeated here.

[0054] The device includes: a mobile phone, a digital camera, a tablet computer, or other devices with a camera function, or a device with an image processing function, or a device with an image display function. The device may include components such as a memory, a processor, an input unit, a display unit, and a power supply.

[0055] Among them, the memory can be used to store software programs and modules, and the processor executes various functional applications and data processing by running the software programs and modules stored in the memory. The memory may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function (such as an image playback function, etc.), etc.; the data storage area may store data created according to the use of the device, etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices. Accordingly, the memory may also include a memory controller to provide the processor and the input unit with access to the memory.

[0056] The input unit can be used to receive input digital or character or image information, and generate keyboard, mouse, joystick, optical or trackball signal input related to user settings and function control. Specifically, the input unit of this embodiment includes not only a camera, but also a touch-sensitive surface (such as a touch display) and other input devices.

[0057] The display unit can be used to display information input by the user or information provided to the user and various graphical user interfaces of the device, which can be composed of graphics, text, icons, videos and any combination thereof. The display unit may include a display panel, and optionally, the display panel may be configured in the form of LCD (Liquid Crystal Display), OLED (Organic Light-Emitting Diode), etc. Further, the touch-sensitive surface may cover the display panel, and when the touch-sensitive surface detects a touch operation on or near it, it is transmitted to the processor to determine the type of touch event, and then the processor provides a corresponding visual output on the display panel according to the type of touch event.

[0058] The embodiment of the present invention further provides a computer-readable storage medium, which may be a computer-readable storage medium included in the memory in the above embodiment; or a computer-readable storage medium that exists independently and is not installed in a device. The computer-readable storage medium stores at least one instruction, which is loaded and executed by a processor to implement Figure 1 The multi-agent path planning method shown. The computer-readable storage medium may be a read-only memory, a disk or an optical disk, etc.

[0059] The embodiment of the present invention also provides a computer program product, including a computer program / instruction, which is loaded and executed by a processor to implement Figure 1 A multi-agent path planning method is shown.

[0060] It should be noted that each embodiment in this specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same or similar parts between the embodiments can be referred to each other. For the device embodiment, equipment embodiment and storage medium embodiment, since they are basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0061] Furthermore, in this document, the terms "comprises," "comprising," or any other variation thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or device that includes the element.

[0062] The above description shows and describes the preferred embodiments of the present invention. It should be understood that the present invention is not limited to the form disclosed herein, and should not be regarded as excluding other embodiments, but can be used in various other combinations, modifications and modifications within the scope of the invention herein, through the above teachings or the technology or knowledge of the relevant field. The changes and modifications made by those skilled in the art do not depart from the spirit and scope of the present invention, and should be within the scope of protection of the claims attached to the present invention.

Claims

1. A multi-agent path planning method, characterized in that: The method comprises: Divide the map into first grids of a preset size, and perform global planning for all agents on the map to generate a global path, wherein the global path is a section from a starting point to a target point that passes through multiple first grids; Performing local planning for all agents based on each path segment of the global path to generate a local path, wherein each path segment is a segment formed by the center points of two adjacent first grids; By real-time monitoring of changes in dynamic obstacles and dynamic targets in the map, real-time data is obtained, and the local path and / or the local path are updated and planned according to the real-time data to obtain an updated path.

2. A multi-agent path planning method according to claim 1, characterized in that: The planning process of the global planning / the local planning includes: Use the Hybird A-star algorithm to perform path planning and obtain the initial path; Performing conflict detection on the initial path, and performing conflict resolution on the detected path conflict through a pre-built conflict tree to obtain a conflict-free path; The global path / the local path is obtained by recursively calling the CBS algorithm to perform repeated conflict detection and conflict processing on the conflict-free path.

3. A multi-agent path planning method according to claim 2, characterized in that: The conflict tree construction process includes: Initialize the root node of the conflict tree, where the root node includes the initial paths of all agents; When a path conflict is detected, the path conflict is added to a conflict node in a conflict tree, wherein each conflict node represents a planning problem of a path segment.

4. A multi-agent path planning method according to claim 3, characterized in that: The step of performing conflict handling on the detected path conflict by using a pre-built conflict tree comprises: Creating two child nodes based on the conflict node in the conflict tree for the detected path conflict to obtain a first child node and a second child node, wherein each child node indicates that an agent in the path conflict is prohibited from appearing at the conflict position; The conflicting position is avoided by re-planning a path for the first agent on the first child node, and the conflicting position is avoided by re-planning a path for the second agent on the second child node.

5. A multi-agent path planning method according to claim 2, characterized in that: Dynamic constraints and resource constraints are introduced into the HybridA-star algorithm to set the cost function.

6. A multi-agent path planning method according to claim 5, characterized in that: The cost function is defined as , where g(x) represents the actual cost from the starting point to the current path point, and h(x) represents the heuristic estimate from the current path point to the target path point. , where ɑ, β, and γ represent the weight of the path length cost PATH_L, the weight of the corner cost TURN_CT, and the weight of the maximum path cost LIMIT_L respectively; ， In the formula, x i ,y i 、z i Indicates the location information of the current path point, x i+1 ,y i+1 、z i+1 Indicates the location information of the next path point; ,in, ， , where represents the corner calculation cost, k represents the weight coefficient, i Indicates the heading angle of the current path point. i+1 Indicates the heading angle of the next path point, R min represents the minimum turning radius, v represents the speed, represents the maximum angular velocity; , where L max Indicates the maximum path length; , x i ,y i 、z i Indicates the location information of the current path point, x goal ,y goal 、z goal Indicates the location information of the target waypoint.

7. A multi-agent path planning method according to claim 1, characterized in that: The updating and planning of the local path and / or the local path according to the real-time data includes: Determine the location of the path conflict according to the real-time data, and if the path conflict is located on the global path, update the planning of the global path and the local path; If the path conflict is on the local path, the local path is updated and planned.

8. A multi-agent path planning device, characterized in that: The device comprises: A global planning unit, used to divide the map into first grids of a preset size, and perform global planning for all agents on the map to generate a global path, wherein the global path is a section from a starting point to a target point passing through a plurality of the first grids; A local planning unit, configured to perform local planning for all agents based on each path segment of the global path to generate a local path, wherein each path segment is a path segment formed by the center points of two adjacent first grids; The path updating unit is used to obtain real-time data by monitoring the changes of dynamic obstacles and dynamic targets in the map in real time, and to update and plan the local path or / and the local path according to the real-time data to obtain an updated path.

9. A multi-agent path planning device, characterized in that: It comprises a processor, a memory and a computer program stored in the memory, wherein the computer program is executed by the processor to implement the steps of a multi-agent path planning method as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program / instructions, wherein the computer program / instructions are executed by a processor to implement the steps of a multi-agent path planning method as described in any one of claims 1 to 7.

Images