Path optimization method and device and storage medium

Abstract

The invention discloses a path optimization method and device and a storage medium, and is applied to multi-agent path planning, and the method comprises the steps: obtaining a passing priority order of a plurality of agents at each path node; selecting a to-be-issued path segment from the planned path of the target agent, and obtaining a traffic priority sequence of target path nodes; selecting at least one agent of which the priority is higher than that of the target agent as a high-priority agent according to the passing priority sequence of the target path node, and obtaining the path conflict cost between the target agent and the high-priority agent; and based on the path conflict cost, determining whether the target path node is issued to the target agent against the passing priority order, so that the target agent preferentially passes through the target path node. In this way, the path execution efficiency can be improved.

Description

Technical Field

[0001] This application relates to the field of path planning technology, specifically to a path optimization method, device, and storage medium. Background Technology

[0002] With the development of robotics technology, the collaborative application of multiple robots plays a key role in various industries such as intelligent manufacturing, express logistics, and service industries, as well as in various aspects such as intelligent warehousing, smart logistics, and automated production.

[0003] Multi-Agent Path Finding (MAPF), a key technology for solving multi-robot path planning, focuses on planning collision-free paths for multiple agents, ensuring they can simultaneously and successfully reach their destination from their respective starting points. For example, in a factory scenario, the factory layout can be abstracted as a directed graph. Each agent has a known starting point and a task point, and each action takes one unit of time. The goal is to plan a sequentially collision-free path for each agent.

[0004] However, there are various uncertainties when agents execute paths, such as different execution speeds, different completion times for each action, different distances between nodes, and the need for robots to accelerate and decelerate to avoid collisions. Therefore, if the execution is strictly carried out according to the MAPF planning sequence, some agents may have to wait because agents with higher priority have not yet passed, resulting in low task execution efficiency. Summary of the Invention

[0005] To address the aforementioned technical problems, this application provides a path optimization method, device, and storage medium to at least solve the problem in related technologies where, if the MAPF planning sequence is strictly followed, some agents may have to wait because agents with higher time priorities have not yet passed through, resulting in low task execution efficiency.

[0006] According to an embodiment of the present invention, a path optimization method is provided, comprising: Obtain the passage priority order of multiple agents at each path node, wherein the passage priority order is the passage order of several agents passing through the path node in the planned path of each agent at the path node. Select the path segment to be distributed from the planned path of the target agent, and obtain the passage priority order of the target path nodes, wherein the target path nodes are the path nodes contained in the path segment to be distributed; According to the passage priority order of the target path nodes, at least one agent with a higher priority than the target agent is selected as a high-priority agent, and the path conflict cost between the target agent and the high-priority agent is obtained. Based on the path conflict cost, it is determined whether to issue the target path node to the target agent in violation of the passage priority order, so that the target agent can pass through the target path node first.

[0007] To solve the above-mentioned technical problems, one technical solution adopted in this application is to provide an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and when the computer program is executed by the processor, it is used to implement the path optimization method in the above-mentioned technical solution.

[0008] To solve the above-mentioned technical problems, one technical solution adopted in this application is to provide a computer-readable storage medium for storing a computer program, which, when executed by a processor, is used to implement the path optimization method in the above-mentioned technical solution.

[0009] The beneficial effects of this application through the above scheme are as follows: The path optimization method provided by this application obtains the passage priority order of multiple agents at each path node, selects the path segment to be sent from the planned path of the target agent, obtains the passage priority order of the target path nodes contained in the path segment to be sent, selects at least one agent with a higher priority than the target agent as a high-priority agent according to the passage priority order of the target path nodes, and obtains the path conflict cost between the target agent and the high-priority agent. Based on the path conflict cost, it is determined whether to send the target path node to the target agent in violation of the passage priority order so that the target agent can pass through the target path node first. In this way, this application determines the high-priority agent based on the passage priority order of the target path nodes, and dynamically judges the path node to be sent to the target agent by obtaining the path conflict cost between the target agent and the high-priority agent, determines whether the target agent can violate the passage priority order, and thus optimizes the timing of the target agent's execution of the planned path when the target agent can violate the passage priority order, thereby improving the path execution efficiency of multiple agents. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1 This is a flowchart illustrating an embodiment of the path optimization method provided in this application; Figure 2 This is a schematic diagram of a single point of conflict provided in this application; Figure 3 This is a schematic diagram of a pair of opposing forces provided in this application; Figure 4 This is a flowchart illustrating another embodiment of the path optimization method provided in this application; Figure 5 This is a schematic diagram of the structure of an embodiment of the electronic device provided in this application; Figure 6 This is a schematic diagram of an embodiment of the computer-readable storage medium provided in this application. Detailed Implementation

[0011] The present application will now be described in further detail with reference to the accompanying drawings and embodiments. It should be particularly noted that the following embodiments are for illustrative purposes only and do not limit the scope of the application. Similarly, the following embodiments are only some, not all, embodiments of the present application, and all other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of the present application.

[0012] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0013] It should be noted that the terms "first," "second," etc., used in this application are for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined with "first," "second," etc., may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0014] Please see Figure 1 , Figure 1 This is a flowchart illustrating an embodiment of the path optimization method provided in this application. It should be noted that if substantially the same result is obtained, this embodiment does not necessarily use it in the same way. Figure 1 The illustrated process sequence is limited. For example... Figure 1 As shown, this embodiment includes: S110: Obtain the passage priority order of multiple agents at each path node.

[0015] An intelligent agent is an execution unit with autonomous mobility that can move along a planned path according to a preset task, such as an automated guided vehicle (AGV) or a mobile robot. A path node is a key location point in the planned path of an intelligent agent; it is a coordinate point that the intelligent agent needs to pass through or stop at during its movement, such as a shelf location point in a factory or a passageway turning point in a logistics warehouse. The passage priority order is the order in which several intelligent agents pass through path nodes in each agent's planned path, used to standardize the passage logic of multiple intelligent agents at nodes.

[0016] In one embodiment, a set of planned paths for several agents is obtained. Based on the temporal information of the planned paths set, the passage order of each agent through each path node is determined to obtain the passage priority order of each path node. The planned paths set contains datasets of the planned paths for each of the several agents, and each agent's planned path consists of consecutive path nodes.

[0017] In one implementation, planned paths of several agents in a target scenario are collected and integrated to form a set of planned paths. This set includes all path nodes in each agent's planned path, as well as the timing information of each agent's passage through each path node. For each path node, all agents whose planned paths contain that node are selected, and a list of associated agents corresponding to each path node is determined. For the list of agents associated with each path node, the timing information of each agent's passage through that node is extracted, and the associated agents are sorted according to the order of the timing information. Based on the sorting result, the passage priority order of the associated agents at each path node is determined, i.e., the passage priority order of each path node is obtained. For example, assume there are two agents. a 1 and a 2, a The path for 1 is 0A-1B-2C-3D. a The path is 0F-1F-2B-3G, indicating that at time 0... a 1. Starting point is A. a 2. Starting point F, time 1 a From point 1 to point B, a 2. Wait in place, 2 moments. a From point 1 to point C, a From point 1 to point B, and so on. Therefore, the path planned by Multi-Agent Path Planning (MAPF) ensures that all agents will not collide in space and time. That is, different agents can pass through the same point, but at different times. According to the MAPF time sequence, the agent with the earlier time of the same path node passes first, and the agent with the later time passes later.

[0018] In other examples, the passing priority order of each path node can also be determined by the estimated passing path node time of each agent. For example, if the estimated time for agent A to reach path node X is t1, and the estimated time for agent B to reach path node X is t2, and t1 < t2, then the passing order of agent A is ranked before that of agent B.

[0019] By obtaining the passing priority order of multiple agents at each path node, it is possible to effectively avoid problems such as collisions and congestion caused by disorderly passing of multiple agents at the same path node, and ensure that the agents move forward stably along the planned path. At the same time, determining the priority based on the timing information of the planned path set ensures that the priority order is consistent with the overall path planning logic of the agents, improving the operation efficiency and coordination of the multi-agent system. S120: Select the path segment to be issued from the planned path of the target agent, and obtain the passing priority order of the target path node.

[0020] The target agent is a specific agent among multiple agents that currently needs to perform path segment issuance and passing priority matching operations. For example, an automated guided vehicle designated to perform warehousing handling tasks. The path execution of the target agent needs to be coordinated with the overall system planning.

[0021] The path segment to be issued refers to the continuous path fragment that is currently selected from the complete planned path of the target agent and needs to be issued to the target agent to drive it to perform the next movement action. It consists of one or more path nodes. The length of the path segment to be issued is dynamically set according to factors such as task progress (such as the number of completed nodes) and scene distance (such as the distance between nodes).

[0022] The target path node is the path node included in the path segment to be issued, which is a subset of the path nodes and is also the key position that the target agent needs to pass through next. It is necessary to confirm its passing priority before issuance to avoid conflicts.

[0023] After confirming the current target agent, a path segment to be deployed is selected from the target agent's planned path. In one embodiment, the complete planned path of the target agent is retrieved from the system database, the path node sequence and timing information in the planned path are parsed, and the current real-time position and progress of the executed path of the target agent are obtained. In this embodiment, a selection rule for the path segment to be deployed can be formulated, which includes at least one of the following: (1) the starting point of the path segment to be deployed is the node where the target agent is currently located or the next node to be executed; (2) the length of the path segment needs to be adapted to the system control cycle. For example, if the control cycle is 5 minutes, then a path segment containing 1-2 path nodes is selected, which is expected to be completed within 5 minutes; (3) avoid marked temporary obstacle nodes, such as nodes in the fault area. The path segment to be deployed is selected from the target agent's planned path according to the selection rule, and the path nodes in the path segment to be deployed are recorded so that they can be used as target path nodes to execute subsequent steps in sequence.

[0024] The system identifies the target path nodes to be evaluated within the path segment to be distributed, and obtains their passage priority order. In one embodiment, if the passage priority order of the current path node is unclear, the planned path set can be retrieved. For the target path node, all agents in the planned paths that contain the target path node are selected, forming a list of associated agents for the target path node. The timing information of each agent passing through the target path node is extracted from the planned path set, and sorted according to the passage order to obtain the sorting result. This sorting result is used as the passage priority order of the target path node. Similarly, the passage priority order of other target path nodes in the path segment to be distributed can be determined and stored in the system control module for later use.

[0025] In this embodiment, the segmentation of the path to be sent can be dynamically adjusted according to the real-time status of the target intelligent agent, avoiding path failure due to scene changes after the complete path is sent, and improving the flexibility of path control.

[0026] S130: Select at least one agent with a higher priority than the target agent as a high-priority agent according to the priority order of the target path nodes, and obtain the path conflict cost between the target agent and the high-priority agent.

[0027] Based on the priority order of the target path nodes, the priority of the target agent passing through the target path nodes is determined. If the target agent itself is the highest priority of the target path node, then there is no need to violate the priority order. The target path node is issued in accordance with the priority order of the target path nodes.

[0028] If the target agent is not the highest priority of the target path node, that is, the target agent needs to wait for at least one other agent with a higher priority than the target agent to pass through the target path node before the target agent can pass, the process is to select at least one agent with a higher priority than the target agent as the high-priority agent and proceed with the subsequent steps according to the priority order of the target path nodes.

[0029] After identifying at least one high-priority agent, the path conflict cost between the target agent and the high-priority agents is obtained. In one embodiment, based on the planned paths of the target agent and the high-priority agents, the path conflict type is determined, and the path conflict cost between the target agent and the high-priority agents is determined based on the path conflict type.

[0030] In one implementation, path conflict types include single-point conflicts and opposing conflicts; see [link to relevant documentation]. Figure 2 and Figure 3 , Figure 2 This is a schematic diagram of a single point of conflict provided in this application. Figure 3 This is a schematic diagram of a one-way conflict provided in this application. A single-point conflict refers to a conflict type where the planned paths of the target agent and the high-priority agent only overlap in travel time or conflict in order at a single target path node, and there are no conflicts on other path segments outside that node, such as... Figure 2 Path node A in the diagram. A reciprocal conflict refers to a conflict path between a target agent and a high-priority agent, where multiple path nodes exist (including the target path node). This means that besides the target path node, there are also passage conflicts at other nodes along the conflict path, such as... Figure 3 The path nodes BCD in the path are the path segments they contain.

[0031] If the planned paths of the target agent and the high-priority agent only conflict at the target path node, the path conflict type is determined to be a single-point conflict; if the planned paths of the target agent and the high-priority agent have conflicting paths, the path conflict type is determined to be a reciprocal conflict, wherein the conflicting path includes multiple path nodes, and the multiple path nodes include the target path node.

[0032] In one implementation, in response to a single-point conflict, the first time the target agent waits for the high-priority agent to pass through the target path node is calculated, and this first time is determined as the path conflict cost. For example, the critical time for the high-priority agent to pass through the target path node is extracted, and the originally planned arrival time of the target agent at the target path node is extracted. The difference between the two is calculated as the first time, which is then determined as the path conflict cost. The longer the first time, the higher the path conflict cost; therefore, violating the passage priority order can be considered to reduce the path conflict cost.

[0033] In response to the path conflict type being a reciprocal conflict, the path conflict cost is determined based on the relationship between the current position of the high-priority agent and the conflict path. In one example, if the high-priority agent has not entered the conflict path, and the distance from the high-priority agent to the endpoint of the conflict path is greater than a preset distance threshold, the second time that the target agent waits for the high-priority agent to pass through the conflict path at an avoidable path node is calculated, and the second time is determined as the path conflict cost. For example, obtaining the current position of the high-priority agent and analyzing the positional relationship between the current position and the conflict path can be divided into two sub-scenarios: (1) The high-priority agent has entered the conflict path: it is possible to query whether there are avoidable path nodes in the planned path of the target agent or the high-priority agent, for example Figure 3 As shown, assuming a 1 is a high-priority agent. a 2 is the target agent. There are avoidable nodes, path node A and path node E. Assume that path node C is the target path node. If there are avoidable path nodes, the time for the target agent to wait for the high-priority agent to pass through the target path node C and to completely pass through path node B at avoidable node A can be calculated, as well as the time for the high-priority agent to wait for the target agent to pass through the target path node C and reach path node D at avoidable node E. Finally, the path optimization strategy is determined based on the actual passage efficiency. For example, the passage order with the shorter comprehensive time can be selected to determine whether the target agent is allowed to violate the original passage priority order of the target path nodes. (2) The high-priority agent has not entered the conflict path: the time when the high-priority agent is about to enter the conflict path can be used to determine whether the target agent is allowed to violate the original passage priority order of the target path nodes. Figure 3 Regarding the path, it's possible that the target agent does not violate the priority order of node C and waits at node A. a 1. Passes, or the target agent violates the priority order of node C and waits at node E. a 1. Pass, or allow a 1. Wait in place until the target agent has completely traversed the conflict path before taking control. a 1. Execute the planned path, etc.

[0034] In one embodiment, after determining the path conflict type based on the planned paths of the target agent and the high-priority agent, in response to the path conflict type being a reciprocal conflict, if there are no avoidable path nodes in the high-priority agent's planned path, or if the distance from the high-priority agent to the endpoint of the conflict path is less than a preset distance threshold, it is determined that the passage priority order will not be violated, and the target path node is issued to the target agent. The conflict path includes multiple path nodes, and these multiple path nodes include the target path node.

[0035] S140: Based on the path conflict cost, determine whether the passage priority order is violated and issue the target path node to the target agent.

[0036] The path conflict cost is a quantified value of system loss caused when a target agent violates the priority order of nodes on the target path, such as entering a node before the predetermined order, and causes a path conflict with other associated agents at the node. The path conflict cost can be calculated by combining the conflict probability, conflict consequences, system recovery costs, etc.

[0037] In one embodiment, the path conflict cost is positively correlated with the time it takes for the target agent to wait for a higher-priority agent to pass through the target path node. If the path conflict cost exceeds a preset threshold, the target agent may be waiting too long according to the original passage order. Therefore, the target agent may be allowed to violate the passage priority order when issuing target path nodes, allowing the target agent to pass through the target path node first. If the path conflict cost does not exceed the preset threshold, the path conflict cost is relatively low when the target agent passes through according to the original passage order. In this case, the target path node is issued according to the passage priority order.

[0038] In one embodiment, the path conflict cost can also be represented as the risk score of the target agent violating the priority order when issuing target path nodes. In this case, the higher the path conflict cost, the more necessary it is to follow the original priority order of the target path nodes, or to prohibit the issuance of target path nodes and trigger path adjustment mechanisms. The specific meaning and calculation method of the path conflict cost can be set according to the actual application scenario and are not limited here.

[0039] This embodiment improves the scientific nature and flexibility of decision-making by quantifying the cost of path conflicts. It can adapt to the cost and security requirements of different scenarios, while allowing flexible deployment in specific conflict scenarios. It can avoid task delays caused by strictly following priorities and improve the path execution efficiency of multi-agent systems.

[0040] Furthermore, to avoid affecting the planned paths of other agents after changing the passage order and to ensure that multiple agents can still avoid conflicts in dynamic scenarios, the multi-agent path planning can be re-executed every preset period, and the path optimization method can be executed based on the new planned path.

[0041] Please see Figure 4 , Figure 4 This is a flowchart illustrating another embodiment of the path optimization method provided in this application. This embodiment uses an automated guided vehicle (AGV) as an example for path execution. It should be noted that if substantially the same result is achieved, this embodiment does not necessarily use that method. Figure 4 The illustrated process sequence is limited. For example... Figure 4 As shown, this embodiment includes: Obtain a set of planned paths for multiple AGVs, and use the MAPF algorithm for multi-vehicle path planning. The input of MAPF is a topology map and multiple AGVs. Each AGV contains a start point and an end point. The output is the set of collision-free paths from the start point to the end point of the AGV, i.e., the planned paths.

[0042] Obtain the node-passage priority order table. Since the path planned by MAPF meets the time-series collision-free requirement, maintain the priority of the AGV passing through the path nodes based on the time sequence of the MAPF path to obtain the passage priority order table for each node. For example, assume there are two vehicles. a 1 and a 2, a The path for car 1 is 0A-1B-2C-3D. a The path for car 2 is 0E-1F-2F-3C-4K. In the point mass model, the priority for point C is... a 1 car higher a Two vehicles strictly followed the path and timing planned by MAPF. a Before car 1 passes point C, a Vehicle 2 cannot pass through this point first. Based on the path planned by MAPF, determine the priority order of each vehicle passing through each path node involved in the path, with vehicles passing through earlier path nodes having higher priority.

[0043] For each AGV that wants to issue a segment of the path, based on the point-passage priority order table, each path node in the segment of the path that each AGV wants to issue is sequentially checked to see if it is allowed to issue the path node according to the timing of the MAPF plan. If the AGV with the highest priority for the path node to be issued is itself, then the path node is issued according to the timing of the MAPF plan. However, since there is often a significant error between the actual path execution and the ideal path planning, these errors may be caused by factors such as different vehicle speeds, acceleration and deceleration, and differences between the actual time occupied by each action and the theoretical calculation. Therefore, during execution, there may be... a When car 1 reaches the 10th moment of the planned path, a Car 2 is still at the point of time 3. At this time, strictly adhering to the timing of the MAPF plan may cause some cars to experience long and unnecessary waits. Therefore, in this embodiment, in order to improve task execution efficiency, under certain circumstances, when the highest priority car of the path node to be sent is not itself, it is allowed to send a segment of the path that violates the timing of the MAPF path.

[0044] For the currently operating AGV vehicles a j , obtain a jThe planned path is obtained by segmenting the path into segments, and then each path node in the pre-deployed segment path is identified as a target path node. It is then determined whether the highest priority of the current target path node is [the target path node]. a j If so, then a j Target path nodes are issued according to their priority order. If the highest priority of the current target path node is not... a j Then, query the priority order table of the target path nodes to find the node with a priority higher than the target path node. a j All cars a k And iterate through them one by one to make further judgments. a j and a k Does it belong to the single point of conflict type?

[0045] In response to a j and a k This is a single-point conflict type, which is allowed under certain circumstances. a j The target path node was assigned in violation of the traffic priority order. For an example, please refer to [link to relevant documentation]. Figure 2 According to the diagram, a 1 car and a Both cars need to pass through point A. The path planned by MAPF is: a The two cars are waiting at the point one step before point A. a Car 1 will continue after passing point A, but if a Car 2 moved faster and arrived at the point before point A much earlier. a The vehicle is moving much slower than expected. If the planned sequence of events is strictly followed... a The two cars have to wait a long time, which is affecting... a The path execution efficiency is 2. Therefore, for single-point conflict types, if the error between the actual execution of the AGV and the ideal planning sequence is not large, such as in this example... a 1 car and a Both vehicles have reached positions very close to point A and both want to issue a waypoint A. Therefore, priority is determined based on timing, allowing... a Vehicle 1 goes first; if the actual execution differs significantly from the plan, then violating the priority order at point A is permissible, allowing... a Two vehicles should proceed first to improve overall traffic efficiency.

[0046] In response to a j anda k This does not belong to the single point of conflict type; further judgment is needed. a k Are there any avoidable nodes in the subsequent planned path, or is the distance to the endpoint of the conflicting path less than a preset distance threshold? If not, then allow... a j If the priority order of passage is violated, the target path node will be issued. If so, the issuance of the current target path node will not be allowed, and the pre-issued path segment will be truncated. The path node that is currently confirmed to have no conflict with other AGVs will be issued, and the process will wait for the judgment of the path segment to be issued in the next round.

[0047] For example, please continue reading Figure 3 , a 1 car and a There is a conflict between the two vehicles traveling in opposite directions on a long passageway. The original planned route is... a Two vehicles yield at path node A and wait. a After vehicle 1 passes, proceed from point A towards BCD. If the actual path is executed... a Car 2 arrived at path node A earlier than the expected arrival time, while a Car 1 has now entered the long passage conflict path, that is a There are no points to avoid between the last point assigned to vehicle 1 and point B. If the path has already been locked to point C, then... a Both vehicles must strictly adhere to the time sequence and continue waiting at point A. They are not allowed to enter the long, conflicting path; otherwise, a deadlock will occur, and one vehicle must reverse to find a safer position. Assume... a Vehicle 1 has not yet entered the conflict path of the long corridor, but is very close to it, such as at point D, where regulations also prohibit it. a Two vehicles violated priority. Only when... a If vehicle 1 has not yet entered the long passage and is more than a certain threshold away from the long passage, if a According to the timing plan of MAPF, the two cars need to wait too long at point A. In this case, it is permissible to allow them to wait. a Two vehicles violated the priority order by entering the lane, and were deemed... a 1. Determine whether to control based on the vehicle's current position and speed. a One vehicle should remain stationary to avoid a deadlock.

[0048] Determine whether all path nodes in the pre-deployment segment path have been evaluated. If not, take the next path node of the current target path node as the current target path node, and continue until all path nodes in the pre-deployment segment path have been evaluated, then end the current path optimization step.

[0049] It should be noted that this solution is based on a lifelong rolling planning framework. Under this framework, the planned path for each robot is not fixed; under certain conditions (such as changes in environmental obstacles or adjustments to task priorities), a new multi-agent path planning (MAPH) cycle will be triggered. Therefore, even... a The two vehicles entering the long, conflicting path disrupted the timing of subsequent paths. However, during the next replanning process, the system will replan the global path for each vehicle based on their real-time new positions. Within the replanned path, a more reasonable adjusted solution may emerge, such as control... a Vehicle 1 should temporarily wait at yield point E to ensure safety. a Two vehicles can pass with priority, ultimately achieving time-series adaptation and conflict avoidance of the overall path.

[0050] Please see Figure 5 , Figure 5 This is a schematic diagram of the structure of an embodiment of the electronic device provided in this application. The electronic device 60 includes a memory 61 and a processor 62 that are interconnected. The memory 61 is used to store a computer program. When the computer program is executed by the processor 62, it is used to implement the path optimization method in the above embodiment.

[0051] The methods described in the above embodiments can exist in the form of a computer program; therefore, this application proposes a computer-readable storage medium. Please refer to [link / reference needed]. Figure 6 , Figure 6 This is a schematic diagram of an embodiment of a computer-readable storage medium provided in this application. The computer-readable storage medium 80 is used to store a computer program 81, which can be executed to implement the path optimization method in the above embodiment.

[0052] The computer-readable storage medium 80 can be any medium capable of storing program code, such as a server, USB flash drive, external hard drive, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0053] The above description is merely an embodiment of this application and does not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A path optimization method, characterized in that, The method, applied to multi-agent path planning, includes: Obtain the passage priority order of multiple agents at each path node, wherein the passage priority order is the passage order of several agents passing through the path node in the planned path of each agent at the path node. Select the path segment to be distributed from the planned path of the target agent, and obtain the passage priority order of the target path nodes, wherein the target path nodes are the path nodes contained in the path segment to be distributed; According to the passage priority order of the target path nodes, at least one agent with a higher priority than the target agent is selected as a high-priority agent, and the path conflict cost between the target agent and the high-priority agent is obtained. Based on the path conflict cost, it is determined whether to issue the target path node to the target agent in violation of the passage priority order, so that the target agent can pass through the target path node first.

2. The method according to claim 1, characterized in that, The path conflict cost is positively correlated with the time the target agent waits for the high-priority agent to pass through the target path node; And / or, determining whether to issue the target path node to the target agent based on the path conflict cost, thereby enabling the target agent to pass through the target path node preferentially, includes: In response to the path conflict cost exceeding a preset threshold, the target path node is sent to the target agent in violation of the passage priority order, so that the target agent can pass through the target path node first.

3. The method according to claim 1 or 2, characterized in that, The cost of obtaining the path conflict between the target agent and the high-priority agent includes: Based on the planned paths of the target agent and the high-priority agent, determine the path conflict type; The path conflict cost between the target agent and the high-priority agent is determined based on the path conflict type.

4. The method according to claim 3, characterized in that, The path conflict types include single-point conflicts and reciprocal conflicts. Determining the path conflict type based on the planned paths of the target agent and the high-priority agent includes: In response to the fact that the planned paths of the target agent and the high-priority agent only conflict at the target path nodes, the path conflict type is determined to be a single-point conflict. In response to a conflict between the planned paths of the target agent and the high-priority agent, the path conflict type is determined to be a reciprocal conflict. The conflict path includes multiple path nodes, and the multiple path nodes include the target path node.

5. The method according to claim 3, characterized in that, The path conflict types include single-point conflicts and opposing conflicts. Determining the path conflict cost between the target agent and the high-priority agent based on the path conflict type includes: In response to the path conflict type being the single-point conflict, the first time the target agent waits for the high-priority agent to pass through the target path node is calculated, and the first time is determined as the path conflict cost. In response to the path conflict type being the opposing conflict, the path conflict cost is determined based on the relationship between the current position of the high-priority agent and the conflict path.

6. The method according to claim 5, characterized in that, The determination of the path conflict cost based on the relationship between the current position of the high-priority agent and the conflict path includes: In response to the high-priority agent not entering the conflict path and the distance from the high-priority agent to the endpoint of the conflict path being greater than a preset distance threshold, the target agent is calculated to wait for the high-priority agent to pass through the conflict path at an avoidable path node for a second time, and the second time is determined as the path conflict cost.

7. The method according to claim 3, characterized in that, After determining the path conflict type based on the planned paths of the target agent and the high-priority agent, the method further includes: In response to the path conflict type being a reciprocal conflict, if there are no avoidable path nodes in the high-priority agent's planned path, or if the distance from the high-priority agent to the endpoint of the conflict path is less than a preset distance threshold, the target path node is sent to the target agent without violating the passage priority order. The conflict path includes multiple path nodes, and the multiple path nodes include the target path node.

8. The method according to claim 1, characterized in that, The step of selecting at least one agent with a higher priority than the target agent as a high-priority agent according to the passage priority order of the target path nodes is performed when the target agent is not the highest priority of the target path node; And / or, the method further includes: In response to the target agent having the highest priority as the target path node, the target path node is issued according to the passage priority order of the target path nodes.

9. The method according to claim 1, characterized in that, The step of obtaining the passage priority order of multiple agents at each path node includes: Obtain a set of planned paths for several of the intelligent agents, wherein the set of planned paths includes several path nodes; Based on the time sequence information of the planned path set, the passage order of each agent through each path node is determined to obtain the passage priority order of each path node; And / or, the method further includes: The path planning of the multiple agents is re-executed every preset period, and the path optimization method is re-performed.

10. An electronic device, characterized in that, The electronic device includes a processor and a memory, the processor being coupled to the memory, the processor being configured to execute one or more steps of the path optimization method according to any one of claims 1 to 9 based on instructions stored in the memory.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that is executed by a processor to implement the steps of the path optimization method as described in any one of claims 1 to 9.

Images