Multi-agent path planning method and system based on priority

Abstract

The invention relates to the technical field of path planning, in particular to a priority-based multi-agent path planning method and system, and the method comprises the steps: abstracting a target scene into an undirected graph; setting a set of agents participating in path planning in the target scene; setting a starting point and an ending point of each agent; defining a priority, a calculation formula of the priority, an initialization rule of a definition and an updating rule in a subsequent path planning step for each agent, wherein the corridor priority represents the defined agent, and the movement priority represents the defined agent; and performing path planning on the agents in the set based on the graph, the starting point and the ending point of each agent, the priority and the calculation formula of the priority. The method is used for improving the quality of the solution generated in the map with the corridor structure.

Description

Technical Field

[0001] This invention relates to the field of intelligent agent path planning technology, and specifically to a priority-based multi-agent path planning method and system. Background Technology

[0002] Multi-agent pathfinding (MAPF) is a problem that involves planning the travel paths of all Automated Guided Vehicles (AGVs) in a target scenario from their respective starting points to their respective destinations, ensuring that no AGV's path collidees with another's while minimizing the AGV's travel time. Currently, MAPF is widely used in automated warehouses, transportation, and parking scenarios.

[0003] However, as the number of AGVs and the travel distance increase, the path space for MAPF search grows exponentially. MAPF has been proven to be NP-hard even when restricted to a two-dimensional mesh graph.

[0004] To improve the efficiency of MAPF search, priority-based planning methods are often used (such as the PIBT (Priority Inheritance with Backtracking) algorithm and the LaCAM (Lazy Constraints Addition Search for MAPF) multi-agent path planning method). These methods plan the movement schemes of AGVs step by step according to their priority from high to low. During the process, low-priority AGVs need to avoid high-priority AGVs to avoid conflicts, thus providing a feasible solution. These algorithms can efficiently solve various MAPF problems, including complex scenarios with large maps and a large number of AGVs.

[0005] However, in maps with corridor structures, especially those with a high density of corridor structures, AGV congestion is likely to occur because there is only one passable road in the corridor, resulting in inconsistent quality of MAPF solutions and room for improvement. Summary of the Invention

[0006] To address the above shortcomings, this invention provides a priority-based multi-agent path planning method and system to solve at least one of the aforementioned problems.

[0007] In a first aspect, the present invention provides a priority-based multi-agent path planning method, the method comprising: Abstract the target scene as an undirected graph. ,in For a set of nodes, Let the graph be a set of edges. It has a corridor; Set the set of intelligent agents participating in path planning in the target scenario. , This represents the total number of agents participating in multi-agent path planning in the target scenario; Configure each agent The starting point and the ending point are denoted as follows: , ,in ; For each intelligent agent Define priority and define priorities Calculation formula and definition , Initialization rules, definition , The update rules in subsequent path planning steps, among which Representative of the defined intelligent agent Corridor priority, Representative of the defined intelligent agent Movement priority; Based on undirected graph Each intelligent agent The starting point and the end point and the defined intelligent agent priority and its calculation formula, for sets The intelligent agent in the process performs path planning.

[0008] Furthermore, The initialization rule is: Initialize =0; The initialization rule is: Initialize ,in Represents nodes in an undirected graph G , The shortest path distance between them without considering other agents.

[0009] Furthermore, in the case of sets During the path planning process of the agents in the system, after each new layout is generated, the process is performed on each agent. Update its movement priority and corridor priority Then use the updated Updated and the update priority of the calculation formula .

[0010] Furthermore, each update , The methods are all: for Determine the intelligent agent Has the destination been reached? If so, calculate... Then update with the calculated result value. The value; if not, then The value increases by 1; for Detecting intelligent agents Is the current location in the corridor? If so, calculate the location of the undirected graph G. and The entrance node of the current corridor Shortest path distance between Then use the calculated shortest path distance renew The value, otherwise will The value is updated to 0; For intelligent agents The node in the new layout.

[0011] Furthermore, the corridor to The entry node is determined according to predefined rules.

[0012] Furthermore, the predefined rule is as follows: for each corridor in the undirected graph G, define its relationship with the agent. The corridor entrance is located at the distance between the intelligent agent and the corridor. The End The node furthest away.

[0013] Furthermore, the target scenario is a warehouse or a maze.

[0014] Secondly, the present invention provides a priority-based multi-agent path planning system, the system comprising: The graph abstraction module is used to abstract the target scene into an undirected graph. ,in For a set of nodes, Let the graph be a set of edges. It has a corridor; The first configuration module is used to set the set of agents participating in path planning in the target scenario. , This represents the total number of agents participating in multi-agent path planning in the target scenario; The second configuration module is used to configure each intelligent agent. The starting point and the ending point are denoted as follows: , ,in ; The third setting module is used to configure each intelligent agent. Define priority and define priorities Calculation formula and definition , Initialization rules, definition , The update rules in subsequent path planning steps, among which Representative of the defined intelligent agent Corridor priority, Representative of the defined intelligent agent Movement priority; The path planning module is used for undirected graph-based paths. Each intelligent agent The starting point and the end point and the defined intelligent agent priority and its calculation formula, for sets The intelligent agent in the process performs path planning.

[0015] Furthermore, in the case of sets During the path planning process of the agents in the system, after each new layout is generated, the process is performed on each agent. Update its movement priority and corridor priority Then use the updated Updated and the update priority of the calculation formula .

[0016] Furthermore, each update , The methods are all: for Determine the intelligent agent Has the destination been reached? If so, calculate... Then update with the calculated result value. The value; if not, then The value increases by 1; for Detecting intelligent agents Is the current location in the corridor? If so, calculate the location of the undirected graph G. and The current corridor opposite Entry node Shortest path distance between Then use the calculated shortest path distance renew The value, otherwise will The value is updated to 0; For intelligent agents The node in the new layout.

[0017] As can be seen from the above technical solutions, the present invention has the following advantages: This invention avoids agent congestion and improves the quality of solutions generated in maps with corridor structures, especially in maps with high corridor density, by using a priority decomposition strategy and a corridor dynamic priority design. Attached Figure Description

[0018] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic flowchart of a method according to an embodiment of the present invention.

[0020] Figure 2 This is a schematic block diagram of a system according to an embodiment of the present invention.

[0021] Figure 3 This is a thumbnail diagram of the maze map maze-128-128-1.

[0022] Figure 4 This is a thumbnail diagram of the warehouse map, warehouse-10-20-10-2-1. Detailed Implementation

[0023] In this embodiment, the intelligent agent is an AGV.

[0024] The key terms used in this invention will be explained below.

[0025] (I) About LaCAM LaCAM, Lazy Constraints Addition Search for MAPF.

[0026] The core structure of the LaCAM algorithm consists of two search layers: the upper search layer is responsible for global search, and the lower search layer is responsible for local planning. The two work together to achieve efficient solution.

[0027] (1) Upper-level search The upper-level search of the LaCAM algorithm is similar to that of the A* algorithm. It unfolds in a multi-dimensional joint state space, using a stack structure (OPEN) to manage upper-level nodes. It explores the state space following a modified depth-first search (DFS) paradigm. The same upper-level node can be visited multiple times. Each time it is visited, the algorithm dynamically adjusts the constraints stored in the node to expand to different successor nodes. A node is only removed from OPEN after all possible successor states have been explored; this approach ensures the completeness of the LaCAM algorithm.

[0028] Each upper-level node Save the following information: Configuration: Records the position information of all AGVs in the undirected graph G at the corresponding time. ,in Represents intelligent agents The node at its current position in the layout.

[0029] Priority array order: order[ ] is a non-negative real number, representing Priority.

[0030] parent pointer: record The parent node is used to maintain path backtracking information.

[0031] Dynamic constraint tree: A tree structure is used to manage the movement constraints of (some) AGVs, while the movement of the remaining AGVs is determined by the lower-level search (PIBT).

[0032] The management of constraint trees follows these principles: Initially, it only contains the root node, which does not contain any constraints.

[0033] The non-root nodes in each layer represent the movement constraints of an AGV. A movement constraint specifies that an AGV must move to a certain node in the next time step.

[0034] The constraint tree is traversed and expanded using a breadth-first search strategy, and maintained using a queue structure. Each time the current parent node is visited, the following constraint tree-related operations are performed: Extract the head constraint node; if the depth of the constraint tree is less than 10 ... (If the root node depth is specified as 0), then a new constraint is generated based on the next layer of AGV of the current node, and it is added to the tail of the queue.

[0035] Collect all constraints on the path from the selected node to the root node and input them into the lower-level search.

[0036] (2) Lower-level search The lower-level search is responsible for "expanding" the upper-level nodes, generating feasible successor states that satisfy the constraints provided by the current upper-level nodes using the PIBT algorithm. LaCAM uses the PIBT algorithm to implement this task, performing fast local path planning.

[0037] LaCAM's upper-level search, through depth-first search and constraint tree expansion, covers the entire solution space, ensuring solution completeness. LaCAM's lower-level search, using the PIBT algorithm, achieves fast local layout planning, guaranteeing both computational efficiency and basic solution quality. The combination of these two approaches enables LaCAM to achieve both solution capability and efficiency when dealing with large-scale MAPF problems.

[0038] (II) PIBT Algorithm PIBT (Priority Inheritance with Backtracking) was originally designed for iteratively solving the MAPF problem. It is a layout generator that starts with an initial layout ( As input, generate a feasible adjacent layout ( By iteratively executing this process, PIBT can quickly generate complete feasible solutions to the MAPF problem.

[0039] The PIBT algorithm is implemented based on a dynamic priority mechanism, as well as priority inheritance and backtracking strategies.

[0040] (1) Dynamic priority mechanism PIBT prioritizes AGVs that have not yet reached the target node or are far from the target node through a dynamic priority mechanism. (Using non-negative real numbers) Represents intelligent agents Priority.

[0041] The specific implementation is as follows: ①Priority initialization: For each Set its initial priority to .

[0042] in, Represents a node , The shortest distance between them This represents the total number of AGVs.

[0043] node , For intelligent agents The starting point and the ending point, the ending point being the target node.

[0044] ②Dynamic adjustment: New layout After generation, for each Adjust its priority as follows: like If the goal is reached, then At this time Set the priority to the lowest level, that is, use Update the calculation results The value (i.e.) ). For intelligent agents New layout The node in which it is located.

[0045] like If the target is not reached, its priority will be increased, that is... The value increases by 1 (i.e. ).

[0046] This design is particularly useful in lifelong MAPF problems (where AGVs are immediately reassigned to a new target location after reaching one). In lifelong scenarios, not all AGVs are at the target location simultaneously; dynamic prioritization ensures task continuity and efficiency.

[0047] ③ Priority inheritance and backtracking strategy: PIBT uses a priority inheritance mechanism to avoid deadlocks during the planning process according to the AGV priority from high to low, and finally generates a feasible solution.

[0048] In the specific implementation of the algorithm, the priority inheritance mechanism is combined with recursion and backtracking strategies to provide opportunities to escape deadlock and seek other feasible solutions.

[0049] The present invention is described in detail below. Specific details, such as particular system structures and techniques, are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of the invention. However, those skilled in the art will understand that the invention may be implemented in other embodiments without these specific details.

[0050] It should be understood that, when used in this specification, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or collections thereof. The terms "comprising," "including," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0051] The terms "one embodiment" or "some embodiments" used in this invention mean that one or more embodiments of the invention include the specific features, structures, or characteristics described in that embodiment. Therefore, phrases such as "in some embodiments" or "in other embodiments" appearing in different parts of the invention do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized.

[0052] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] Figure 1 This is a flowchart illustrating a priority-based multi-agent path planning method provided in an embodiment of the present invention. Figure 1 The executing entity can be a priority-based multi-agent path planning system. The multi-agent path planning method provided in this embodiment is executed by a computer device; correspondingly, the priority-based multi-agent path planning system runs on the computer device. Depending on different requirements, the order of steps in this flowchart can be changed, and some steps can be omitted.

[0054] Please refer to Figure 1 The method includes steps 101 to 105.

[0055] Step 101: Abstract the target scene into an undirected graph. ,in For a set of nodes, Let the graph be a set of edges. It has a corridor; Step 102: Set the set of agents participating in path planning in the target scenario. , This represents the total number of agents participating in multi-agent path planning in the target scenario; Step 103: Configure each agent The starting point and the ending point are denoted as follows: , ,in ; Step 104: For each intelligent agent Define priority and define priorities Calculation formula and definition , Initialization rules, definition , The update rules in subsequent path planning steps, among which Representative of the defined intelligent agent Corridor priority, Representative of the defined intelligent agent Movement priority; Step 105: Based on the above undirected graph Each intelligent agent The starting point and the end point and the defined priority and its calculation formula For sets The intelligent agent in the process performs path planning.

[0056] In one embodiment, The initialization rules are as follows: =0, The initialization rules are as follows: ,in Represents nodes in an undirected graph G , The shortest path distance between them without considering other agents.

[0057] That is: Initialize according to its initialization rules, and the initial value is 0; Initialize according to its initialization rules to obtain the initial value. .

[0058] Understandably, based on the calculation formula , The initial value is equal to initial value and The sum of the initial values.

[0059] That is initial value .

[0060] Understandably, the present invention initializes the intelligent agent. Movement priority Previously, for each intelligent agent First search its nodes , The shortest path between .

[0061] In another embodiment of the present invention, the method of this embodiment is applied to a set. During the path planning process of the agent in the process, each time a new layout (denoted as ) is generated... After that, for each intelligent agent Update its movement priority and corridor priority Then use the updated , and the calculation formula Update priority .

[0062] Understandably, each time a new layout is generated, for each agent... Update its movement priority and corridor priority Then use the updated version , Sum update priority .

[0063] As an illustrative embodiment of the present invention, each update of the present invention... , The methods are all: for Determine the intelligent agent Has the destination been reached? If so, calculate... Then update with the calculated result value. The value; if not, then The value increases by 1; for Detecting intelligent agents Is the current location in the corridor? If so, calculate the location of the undirected graph G. and The current corridor opposite The entry node (this entry node is denoted as) The shortest path distance between ) Then use the calculated shortest path distance renew The value, otherwise will The value is updated to 0; For intelligent agents In the new layout The node in which it is located.

[0064] In this invention, the definition of "corridor" is as follows: In an undirected graph G, a subgraph is called a corridor if and only if: The subgraph is connected; This subgraph contains, and only contains: a) Several nodes with a degree of 2; b) There are exactly two nodes with a degree greater than 2.

[0065] Based on this definition of corridors, the present invention identifies corridor structures in an undirected graph G.

[0066] Understandably, in a corridor, two nodes with a degree greater than 2 are the corridor's entrance and exit. One is the exit, and the other is the entrance.

[0067] In one embodiment, the invention further includes, before the agent's corridor priority is updated for the first time: Searching for an undirected graph The corridor in the middle yields an undirected graph. Collection of all corridors ,in Representing an undirected graph The total number of central corridors Represents the undirected graph found The first in A corridor, , for The set of nodes, for The set of edges in; .

[0068] Optional, corridor For intelligent agents The entry node is determined according to predefined rules. .

[0069] Understandably, this predefined rule is pre-set in this invention.

[0070] Optionally, the above predefined rule can be set as follows: for each corridor in the undirected graph G Define its relationship with intelligent agents The corridor entrance is: Corridor Mid-range intelligent agents The End The farthest node. Among them... , .

[0071] It should be noted that the shortest paths mentioned in this explanation are the shortest paths without considering other agents, and the shortest path distances are the shortest path distances without considering other agents. For example, calculating the shortest path distance on an undirected graph G... and The entrance node of the current corridor Shortest path distance between That is, computing the undirected graph G without considering other agents. and The entrance node of the current corridor Shortest path distance between .

[0072] In practice, skilled personnel can also reset the above predefined rules according to actual needs. For example, the above predefined rule is: when the intelligent agent... Located in the corridor The At each node, the corridor For intelligent agents The definition rule for the corridor entrance is: obtain the intelligent agent Without considering other intelligent agents, from the corridor The Each node to its path planning endpoint The shortest path, and the corridors located on that shortest path. Nodes with a degree greater than 2 are corridors. The current exit for this intelligent agent, the corridor Another node with a degree greater than 2 is a corridor. The current entry point (i.e., the entry node) for this agent. , Representing the corridor The total number of nodes in the middle. This predefined rule implements each corridor. The method for dynamically defining corridor entrances for different agents supports the scenario where some or all agents "reverse" into the corridor. Understandably, under this predefined rule, the undirected graph G described above... and The current corridor opposite The entry node should be changed to: on the undirected graph G and Current situation The entry node, where represent The corridor we are currently in.

[0073] As an illustrative embodiment of the present invention, the PIBT algorithm is used to generate each new layout.

[0074] The PIBT (Priority Inheritance with Backtracking) algorithm and the LaCAM (LazyConstraints Addition Search for MAPF) algorithm are among the most advanced fast MAPF algorithms currently available, and both belong to the priority planning algorithm class.

[0075] To illustrate the invention more clearly, the multi-agent path planning method of the present invention will be described below based on LaCAM and PIBT (named LaCAMCP (LaCAM with Corridor Priority)).

[0076] In this embodiment, the multi-agent path planning method includes: Step 1050: Input MAPF event; The input MAPF event includes steps 101 to 104 as described above.

[0077] Step 1051: Initialize the OPEN stack.

[0078] Step 1052: Remove the top node of the OPEN stack.

[0079] Step 1053: Determine whether all agents in the top node of the OPEN stack have reached their respective endpoints: If so, then generate a set. The paths of each agent are calculated and output, and the path planning ends. If not, proceed to step 1054.

[0080] Step 1054: Generate a new layout Then proceed to step 1055.

[0081] Step 1055: For each agent Update its priority Then proceed to step 1056.

[0082] Step 1055 specifically includes: for each agent The above update is adopted. , Methods to update intelligent agents Movement priority and corridor priority Then use the updated , and the above calculation formula Update priority .

[0083] Step 1056: Construct a new node and add it to the top of the OPEN stack. Then return to step 1052 to continue execution.

[0084] Step 1055 specifically includes: based on the new layout and various intelligent agents Updated priority Construct a new node and add it to the top of the OPEN stack. Then return to step 1052 to continue execution.

[0085] Based on the above steps, in this embodiment, each agent is updated after a new layout is generated. The agent is then updated based on its movement priority and corridor priority. priority .

[0086] Understandably, steps 1051 to 1056, i.e., based on undirected graphs Each intelligent agent The starting point and the end point and the defined intelligent agent priority and its calculation formula, for sets The intelligent agent in the process performs path planning.

[0087] In this embodiment, the PIBT algorithm is used to generate a new layout. .

[0088] It can be seen that the present invention proposes a new dynamic priority update method for intelligent agents. It not only retains the characteristics of the original dynamic priority, but also assigns higher priority to intelligent agents that are farther away from the target, thereby reducing the overall completion time of the system, and can also solve the problem in the corridor map.

[0089] In another embodiment, initializing the OPEN stack in step 1051 above includes: Constructing a system that includes each intelligent agent The nodes in the initial state are pushed onto the OPEN stack.

[0090] In the method of this embodiment, the intelligent agent The initial states include: intelligent agent The current node position in the undirected graph G; and based on the definition , Initialization rules , Substitute into the calculation formula Initialization priority ,Right now .

[0091] In some other embodiments, the intelligent agent The initial state also includes: Based on definition , Initialization rules , ,Right now =0, .

[0092] Understandably, the newly constructed nodes described above not only contain intelligent agents. Priority after current round optimization It also includes the priority used in the current round to calculate the optimized priority. Dynamically updated , So that it can be used in the next round. Optimize usage.

[0093] As an illustrative embodiment of the present invention, the target scenario is a warehouse or a maze.

[0094] When the target scenario is a warehouse, the physical space within the warehouse can be abstracted as an undirected graph, with the endpoints and intersections of the passageways between shelves that allow agents to pass as nodes and the passageways between these nodes as edges. Therefore, in the undirected graph G, This represents the set of endpoints and intersections of passageways between shelves within the warehouse that allow intelligent agents to pass through. This represents the set of passageways within the warehouse that allow intelligent agents to pass through. Understandably, when the target scenario is a warehouse, this set... The agents in the text are those involved in path planning within the warehouse.

[0095] Please refer to Figure 2 , Figure 3 , Figure 2 Let maze-128-128-1 be an abstract maze map. This maze map is a 128×128 square grid maze (corresponding to "128-128" in maze-128-128-1). In maze-128-128-1: maze is an environment name referring to a maze; the final "1" indicates that the width of the paths within the maze is 1, meaning that AGVs (automatic vehicles) can only pass one at a time, not side-by-side.

[0096] Figure 3 The map for an abstract warehouse is warehouse-10-20-10-2-1. In this map, the warehouse has 20 rows and 10 columns of containers (corresponding to "10-20" in warehouse-10-20-10-2-1). Furthermore, in warehouse-10-20-10-2-1: Warehouse represents the storage environment, "10-2" indicates each container is 10 units long and 2 units wide, and the final "1" indicates that the horizontal and vertical spacing between containers is 1 unit.

[0097] The LaCAMCP and LaCAM-based multi-agent path planning algorithms of this invention were experimentally applied. Figure 2 , Figure 3 Multi-agent path planning was performed on the map shown, with SOC (Sum of Cost) as the objective function. Tables 1 and 2 show the reduced result after dividing SOC by its theoretical lower bound (LB). Table 1 corresponds to maze-128-128-1, and Table 2 corresponds to warehouse-10-20-10-2-1. The experiment set a time limit of 5 seconds for solving a single MAPF problem; exceeding this time limit was considered a failure. Based on this, the success rates of LaCAM and LaCAMCP in maze-128-128-1 were compared (no failures were observed in warehouse-10-20-10-2-1, so this metric was not considered for this map), as well as the computation time for successful solutions on both maps. In Tables 1 and 2, SOC, or sum of cost, represents the total distance traveled by all AGVs. LB is the lower bound of the SOC for the entire problem. A SOC / LB ratio closer to 1 indicates a better algorithm. Computation time is the time taken by the algorithm for path planning. The success rate represents the proportion of successfully calculated cases out of the total number of cases within a specified time.

[0098] Table 1: maze-128-128-1

[0099] Table 2: warehouse-10-20-10-2-1

[0100] Experiments show that LaCAMCP outperforms LaCAM in solution time on both maps, while also improving solution quality. On maze-128-128-1 and warehouse-10-20-10-2-1, the SOC improvement is approximately [missing information]. And 5%. In addition, compared with LaCAM, LaCAMCP has a higher success rate in solving maze-128-128-1, which shows that the present invention improves the solution capability.

[0101] like Figure 4 As shown, this invention provides a priority-based multi-agent path planning system, the system comprising: Graph abstraction module 201 is used to abstract the target scene into an undirected graph. ,in For a set of nodes, Let the graph be a set of edges. It has a corridor; The first setting module 202 is used to set the set of intelligent agents participating in path planning in the target scenario. , This represents the total number of agents participating in multi-agent path planning in the target scenario; The second setting module 203 is used to set each intelligent agent. The starting point and the ending point are denoted as follows: , ,in ; The third setting module 204 is used to set up each intelligent agent. Define priority and define priorities Calculation formula and definition , Initialization rules, definition , The update rules in subsequent path planning steps, among which Representative of the defined intelligent agent Corridor priority, Representative of the defined intelligent agent Movement priority; Path planning module 205 is used to plan the path based on the above undirected graph. Each intelligent agent The starting point and the end point and the defined priority and its calculation formula, for sets The intelligent agent in the process performs path planning.

[0102] In one embodiment, the path planning module 205 performs a set During the path planning process of the agents in the system, after each new layout is generated, the process is performed on each agent. Update its movement priority and corridor priority Then use the updated Updated and the update priority of the calculation formula .

[0103] As an illustrative embodiment of the present invention, each update , The methods are all: for Determine the intelligent agent Has the destination been reached? If so, calculate... Then update with the calculated result value. The value; if not, then The value increases by 1; for Detecting intelligent agents Is the current location in the corridor? If so, calculate the location of the undirected graph G. and The current corridor opposite Entry node Shortest path distance between Then use the calculated shortest path distance renew The value, otherwise will The value is updated to 0; For intelligent agents The node in the new layout.

[0104] This embodiment of the priority-based multi-agent path planning system belongs to the same inventive concept as the multi-agent path planning methods in the above embodiments. For details not described in detail in the embodiment of the priority-based multi-agent path planning system, please refer to the embodiments of the multi-agent path planning methods mentioned above.

[0105] The same or similar parts between the various embodiments in this specification can be referred to mutually. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple, and the relevant parts can be referred to the description in the method embodiments.

[0106] In the embodiments provided by this invention, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0107] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0108] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0109] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0110] Without significantly increasing the time and space complexity of the LaCAM algorithm, the LaCAMCP algorithm significantly optimizes the performance and behavior of the algorithm on corridor maps through priority decomposition strategy and corridor dynamic priority design, including the quality of the generated solutions and the solution speed.

[0111] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A priority-based multi-agent path planning method, characterized in that, The methods include: Abstract the target scene as an undirected graph. ,in For a set of nodes, Let the graph be a set of edges. It has a corridor; Set the set of intelligent agents participating in path planning in the target scenario. , This represents the total number of agents participating in multi-agent path planning in the target scenario; Configure each agent The starting point and the ending point are denoted as follows: , ,in ; For each intelligent agent Define priority and define priorities Calculation formula and definition , Initialization rules, definition , The update rules in subsequent path planning steps, among which Representative of the defined intelligent agent Corridor priority, Representative of the defined intelligent agent Movement priority; Based on undirected graph Each intelligent agent The starting point and the end point and the defined intelligent agent priority and its calculation formula, for sets The intelligent agent in the process performs path planning.

2. The method according to claim 1, characterized in that, The initialization rule is: Initialize =0; The initialization rule is: Initialize ,in Represents nodes in an undirected graph G , The shortest path distance between them without considering other agents.

3. The method according to claim 1, characterized in that, In the set During the path planning process of the agents in the system, after each new layout is generated, the process is performed on each agent. Update its movement priority and corridor priority Then use the updated Updated and the update priority of the calculation formula .

4. The method according to claim 3, characterized in that, Each update , The methods are all: for Determine the intelligent agent Has the destination been reached? If so, calculate... Then update with the calculated result value. The value; if not, then The value increases by 1; for Detecting intelligent agents Is the current location in the corridor? If so, calculate the location of the undirected graph G. and The current corridor opposite Entry node Shortest path distance between Then use the calculated shortest path distance renew The value, otherwise will The value is updated to 0; For intelligent agents The node in the new layout.

5. The method according to claim 4, characterized in that, Corridor The entry node is determined according to predefined rules.

6. The method according to claim 5, characterized in that, The predefined rule is: For each corridor in an undirected graph G, define its relationship to the agent. The corridor entrance is located at the distance between the intelligent agent and the corridor. The End The node furthest away.

7. The method according to any one of claims 1 to 6, characterized in that, The target scenario is a warehouse or a maze.

8. A priority-based multi-agent path planning system, characterized in that, The system includes: The graph abstraction module is used to abstract the target scene into an undirected graph. ,in For a set of nodes, Let the graph be a set of edges. It has a corridor; The first configuration module is used to set the set of agents participating in path planning in the target scenario. , This represents the total number of agents participating in multi-agent path planning in the target scenario; The second configuration module is used to configure each intelligent agent. The starting point and the ending point are denoted as follows: , ,in ; The third setting module is used to configure each intelligent agent. Define priority and define priorities Calculation formula and definition , Initialization rules, definition , The update rules in subsequent path planning steps, among which Representative of the defined intelligent agent Corridor priority, Representative of the defined intelligent agent Movement priority; The path planning module is used for undirected graph-based paths. Each intelligent agent The starting point and the end point and the defined intelligent agent priority and its calculation formula, for sets The intelligent agent in the process performs path planning.

9. The system according to claim 8, characterized in that, In the set During the path planning process of the agents in the system, after each new layout is generated, the process is performed on each agent. Update its movement priority and corridor priority Then use the updated Updated and the update priority of the calculation formula .

10. The system according to claim 9, characterized in that, Each update , The methods are all: for Determine the intelligent agent Has the destination been reached? If so, calculate... Then update with the calculated result value. The value; if not, then The value increases by 1; for Detecting intelligent agents Is the current location in the corridor? If so, calculate the location of the undirected graph G. and The current corridor opposite Entry node Shortest path distance between Then use the calculated shortest path distance renew The value, otherwise will The value is updated to 0; For intelligent agents The node in the new layout.

Images