A path planning method, device, equipment and storage medium

Abstract

This application provides a path planning method, apparatus, device, and storage medium. The method includes: determining an initial planned path for an agent to be planned based on a pre-constructed road adjacency matrix; determining a corresponding three-dimensional time window based on the initial planned path, the three-dimensional time window including the travel time period when each agent passes through each path node in the planned path corresponding to that agent; updating the initial planned path based on a first travel time period when the agent to be planned passes through each path node in the initial planned path and a second travel time period when the planned agent passes through each path node in the corresponding target planned path, to obtain the target planned path for the agent to be planned. This application can achieve collision-free path planning for any agent in a multi-agent task scheduling scenario, reducing resource overhead during multi-agent path planning, avoiding deadlock in multi-agent task scheduling, and improving the safety of multiple agents under target planned paths.

Description

Technical Field

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

[0002] With the rapid development of the Industrial Internet, multiple intelligent agents are typically configured in fields such as warehouse management, logistics scheduling, urban transportation, and manufacturing. These agents include warehouse robots, logistics robots, automated guided vehicles (AGVs), and driverless vehicles. The goal is to schedule these agents to complete the assigned tasks.

[0003] In multi-agent task scheduling scenarios, paths are typically planned according to the tasks assigned to each agent. However, when multiple agents move along the planned paths, collisions may occur, leading to deadlock during the scheduling process and preventing successful task execution. Summary of the Invention

[0004] This application provides a path planning method, apparatus, device, and storage medium that enables collision-free path planning for any agent in a multi-agent task scheduling scenario through a three-dimensional time window, reducing resource overhead during multi-agent path planning and improving the movement safety of any agent on the target planned path.

[0005] In a first aspect, embodiments of this application provide a path planning method, the method comprising:

[0006] In response to the task scheduling instruction of the agent to be planned, the initial planning path of the agent to be planned is determined according to the pre-constructed road adjacency matrix;

[0007] Based on the initial planned path, a corresponding three-dimensional time window is determined. The three-dimensional time window includes the time period when each agent passes through each path node in the planned path corresponding to the agent.

[0008] The initial planning path is updated based on the first travel time when the agent to be planned passes through each path node in the initial planning path and the second travel time when the planned agent passes through each path node in the corresponding target planning path, so as to obtain the target planning path of the agent to be planned.

[0009] Secondly, embodiments of this application provide a path planning device, the device comprising:

[0010] The initial path determination module is used to determine the initial planning path of the agent to be planned based on a pre-constructed road adjacency matrix in response to the task scheduling instruction of the agent to be planned.

[0011] The time window determination module is used to determine the corresponding three-dimensional time window based on the initial planned path. The three-dimensional time window includes the time period when each agent passes through each path node in the planned path corresponding to the agent.

[0012] The path planning module is used to update the initial planning path based on the first travel time when the agent to be planned passes through each path node in the initial planning path and the second travel time when the planned agent passes through each path node in the corresponding target planning path, so as to obtain the target planning path of the agent to be planned.

[0013] Thirdly, embodiments of this application provide an electronic device, which includes:

[0014] A processor and a memory, the memory being used to store a computer program, and the processor being used to call and run the computer program stored in the memory to execute the path planning method provided in the first aspect of this application.

[0015] Fourthly, embodiments of this application provide a computer-readable storage medium for storing a computer program that causes a computer to perform the path planning method provided in the first aspect of this application.

[0016] Fifthly, embodiments of this application provide a computer program product, including a computer program / instructions that, when executed by a processor, implement the path planning method provided in the first aspect of this application.

[0017] This application provides a path planning method, apparatus, device, and storage medium. Responding to a task scheduling instruction from any agent to be planned, the method first determines an initial planned path for the agent based on a pre-constructed road adjacency matrix, thereby determining a corresponding three-dimensional time window. This three-dimensional time window includes the travel time intervals when each agent passes through each path node in its corresponding planned path. Then, based on the first travel time interval when the agent to be planned passes through each path node in the initial planned path and the second travel time interval when a planned agent passes through each path node in its corresponding target planned path, the initial planned path is updated to obtain the target planned path for the agent to be planned. This ensures that there are no collisions between the agent to be planned and any planned agents in the target planned path, thereby achieving collision-free path planning for any agent in a multi-agent task scheduling scenario. This reduces resource overhead during multi-agent path planning, avoids deadlock in multi-agent task scheduling, improves the movement safety of any agent on the target planned path, and ensures the successful execution of the scheduled task by any agent. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments 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 flowchart illustrating a path planning method according to an embodiment of this application;

[0020] Figure 2 This is a schematic diagram of the structure of a three-dimensional time window as shown in an embodiment of this application;

[0021] Figure 3 This is a schematic diagram illustrating the time periods traversed by the agent to be planned at each path node, as shown in an embodiment of this application.

[0022] Figure 4 A flowchart illustrating another path planning method as shown in an embodiment of this application;

[0023] Figure 5 This is a schematic diagram illustrating an embodiment of the present application, showing how to determine whether there is a first-type collision between the agent to be planned and any first-type planned agent at the current path node;

[0024] Figure 6 This is a schematic diagram illustrating how to determine whether there is a second-type collision between the agent to be planned and any second-type planned agent in the current sub-path from the current path node to the next path node, as shown in an embodiment of this application.

[0025] Figure 7 This is a schematic block diagram of a path planning device shown in an embodiment of this application;

[0026] Figure 8 This is a schematic block diagram of an electronic device shown in an embodiment of this application. Detailed Implementation

[0027] 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.

[0028] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. 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 server that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0029] To address the deadlock issue that may arise from collisions between planned paths and other agents in multi-agent task scheduling scenarios, this application proposes a path planning scheme to eliminate potential collisions within the planned path of any agent. A three-dimensional time window structure is used to record the time intervals during which each agent passes through each path node in its planned path. Then, upon receiving a task scheduling instruction from any agent to be planned, the initial planned path for that agent is first determined based on a pre-constructed road adjacency matrix, thereby establishing the corresponding three-dimensional time window. Furthermore, based on the first passage time when the agent to be planned passes through each path node in the initial planned path and the second passage time when the planned agent passes through each path node in the corresponding target planned path, the system analyzes whether the agent to be planned collides with any of the planned agents when passing through each path node. This updates the initial planned path to obtain the target planned path for the agent to be planned, ensuring that there are no collisions between the agent to be planned and any of the planned agents in the target planned path. This achieves collision-free path planning for any agent in a multi-agent task scheduling scenario, improving the motion safety of any agent on the target planned path.

[0030] Figure 1 This is a flowchart illustrating a path planning method according to an embodiment of this application. (Refer to...) Figure 1 The method may include the following steps:

[0031] S110, in response to the task scheduling instruction of the agent to be planned, determines the initial planning path of the agent to be planned based on the pre-constructed road adjacency matrix.

[0032] In this context, intelligent agents can be intelligent entities such as warehouse robots, logistics robots, AGVs, and autonomous vehicles that perform various tasks within a defined specific road scenario. By scheduling these intelligent agents within this specific road scenario to execute their assigned tasks, each intelligent agent can travel on different roads within that scenario.

[0033] In this application, to ensure accurate navigation of multiple agents within a specific road scenario, it is first necessary to pre-analyze the various roads existing in that scenario. Therefore, this application can first identify all intersections within the specific road scenario. At any intersection, an agent may choose a different road. Then, to ensure accurate navigation of multiple agents within the specific road scenario, this application can treat each intersection within the specific road scenario as a corresponding path node, analyze the adjacency relationship between any two path nodes, and thus construct the corresponding road adjacency matrix.

[0034] In some implementations, when assigning a task to any agent, a task scheduling instruction is first generated for that agent. This agent is the agent to be planned in this application, and the task scheduling instruction carries the starting path node currently occupied by the agent when executing the assigned task and the final target path node to be reached.

[0035] In response to the task scheduling instruction of the agent to be planned, the first step is to analyze the traversable path from the starting path node to the target path node of the agent to be planned based on the adjacency relationship between every two path nodes in the pre-constructed road adjacency matrix, and use this as the initial planning path of the agent to be planned.

[0036] As an optional implementation of this application, when a task scheduling instruction for the agent to be planned is detected, this application can determine the initial planning path of the agent to be planned through the following steps:

[0037] The first step is to construct the corresponding road adjacency matrix based on the path nodes in the road scene map where the agent to be planned is located.

[0038] This application can represent the adjacency relationship between two path nodes by the length of the walkable path between them. For example, if there is no walkable path between two path nodes, the path length between them can be set to infinity as the corresponding inaccessible marker. Conversely, if there is a walkable path between two path nodes, indicating that the two path nodes are adjacent, the length of the walkable path between them can be used as the corresponding path length.

[0039] Therefore, this application can first obtain a road scene map of the multi-agent environment, analyze the various intersections in the road scene map as corresponding path nodes, and determine the position coordinates of each path node. Then, by analyzing whether there is a passable road between each pair of path nodes and the length of the passable road, a corresponding road adjacency matrix can be constructed.

[0040] For example, assuming that the road scene map in which the multi-agent is located has N path nodes, the road adjacency matrix can be an N*N two-dimensional matrix, where each matrix element is the path length between the corresponding two path nodes.

[0041] The second step is to search for the shortest path between the starting path node and the target path node of the agent to be planned based on the road adjacency matrix, so as to obtain the initial planning path of the agent to be planned.

[0042] After detecting the task scheduling instruction of the agent to be planned, the road adjacency matrix corresponding to the road scene map where the agent is located is first determined in order to analyze whether there is a passable road between each two path nodes and the length of the passable road.

[0043] When planning a task path for an agent to be planned, the current path node of the agent can be used as the starting path node, and the target path node required for the agent to execute the assigned task can be determined. Then, based on the path length between any two path nodes in the road adjacency matrix, a shortest path algorithm such as Floyd's algorithm or Dijkstra's algorithm can be used to search for the shortest path from the starting path node to the target path node. Furthermore, since the search for this shortest path does not consider whether there are collisions with paths already planned by other agents, this application can use this shortest path as the initial planned path for the agent to be planned. This allows for further optimization of the initial planned path based on path collisions with other planned agents, thereby avoiding collisions between multiple agents during task scheduling.

[0044] S120. Based on the initial planned path, determine the corresponding three-dimensional time window. The three-dimensional time window includes the time period when each agent passes through each path node in the planned path corresponding to the agent.

[0045] Considering that the initial planned path of the agent to be planned is determined without considering collisions with the paths already planned by other agents, and in order to ensure the safe passage of multiple agents during task scheduling, it is necessary to avoid collisions between multiple agents as much as possible. Therefore, this application needs to analyze the travel time of each agent when passing through each path node in the agent's planned path, and the travel time corresponding to each path node can include the entry time and exit time of that path node.

[0046] As can be seen from the above, this application can construct a three-dimensional time window data structure from three dimensions: agent, path node, and time. This allows for the rapid location of the time window of any agent at any path node on a time window model. The time window refers to the time period occupied by any agent from entering any path node to leaving that path node.

[0047] Therefore, in order to avoid collisions among multiple agents during task scheduling, it is required that during the travel time of any agent passing through any path node, other agents are not allowed to pass through that path node, while other agents can pass through the space time outside the travel time occupied by that path node.

[0048] Among them, such as Figure 2 As shown, the data structure of the three-dimensional time window can be a time window matrix constructed from two dimensions: agent and path node. The matrix elements of the time window matrix can be the time window represented by the travel time of any agent under any path node, which is in the form of a linked list. Each time window can include the entry time point and exit time point of any agent under any path node.

[0049] In this application, agents can be categorized into three types: agents awaiting planning, planned agents, and unplanned agents. For unplanned agents, there is no task scheduling requirement, and therefore no travel time segment for each path node. Thus, the travel time segment for an unplanned agent is empty within the three-dimensional time window. For planned agents, however, a corresponding path has already been planned for them. Therefore, the three-dimensional time window stores the travel time segment for each planned agent when passing each path node in the planned path.

[0050] For the agent to be planned, after determining the initial planning path of the agent, the entry time and exit time of the agent when passing through each path node in the initial planning path can be calculated based on the agent's travel speed, agent size, and path length between adjacent path nodes in the initial planning path, thereby obtaining the corresponding travel time period.

[0051] For example, such as Figure 3 As shown, taking the initial planned path of the agent to be planned as including a starting path node, a target path node, and an intermediate path node as an example, assume that the path length between any two adjacent path nodes is D, the length and width of the agent to be planned are L and W respectively, and the travel speed is V. To avoid collisions between any two agents at a certain path node, a square area with the width of the agent as the center can be used as the node area corresponding to the agent's passage through that path node. It is required that no other agent can pass through the path node when the agent to be planned passes through it.

[0052] Therefore, for the starting path node and the target path node, the planning agent will stop at the center of the starting path node and start driving outward. After the planning agent enters the target path node, it will not drive out, but will stop at the center of the target path node.

[0053] Therefore, as Figure 3 As shown, the time consumed by the agent to be planned at the starting path node and the target path node can be t1 = (L+W) / 2V, the time consumed at the intermediate path node can be t2 = (L+W)V, and the path consumption time of the agent to be planned between every two adjacent path nodes can be t3 = (DLW)V.

[0054] The path time of the agent to be planned between any two adjacent path nodes is not fixed and will vary with the path length D between the two path nodes.

[0055] Assuming the task start time of the agent to be planned at the starting path node is T, then the entry time of the agent at the starting path node is T1 = T, and the exit time is T2 = T + t1, resulting in a travel time of (T, T + t1) when the agent passes through the starting path node. The entry time at the intermediate path node is T3 = T2 + t3, and the exit time is T4 = T3 + t2, resulting in a travel time of (T3, T3 + t2) when the agent passes through the intermediate path node. Finally, the entry time at the target path node is T5 = T4 + t3, and the exit time is T6 = T5 + t1, resulting in a travel time of (T5, T5 + t1) when the agent passes through the target path node.

[0056] S130, based on the first travel time when the agent to be planned passes through each path node in the initial planning path and the second travel time when the planned agent passes through each path node in the corresponding target planning path, the initial planning path is updated to obtain the target planning path of the agent to be planned.

[0057] After determining the corresponding three-dimensional time window based on the initial planning path of the agent to be planned, the first travel time when the agent to be planned passes through each path node in the initial planning path, and the second travel time when each planned agent passes through each path node in its pre-planned target planning path can be determined within the three-dimensional time window.

[0058] Based on the above, the time intervals during which the agent to be planned travels between any two adjacent path nodes in the initial planned path, and the time intervals during which each planned agent travels between any two adjacent path nodes in the planned target path can also be determined.

[0059] Then, based on the time periods of the agent to be planned on the initial planned path and the time periods of each planned agent on the planned target path, it is determined whether there is any partial overlap in the time periods of the agent to be planned and any planned agent on a certain road segment, thereby determining whether there is a possibility of collision between the agent to be planned and any planned agent.

[0060] If a collision occurs between the agent to be planned and any of the already planned agents, the initial planned path can be continuously updated by either updating the travel time of the agent at the collision path nodes or directly searching for a new path from the starting path node to the target path node. This process continues until the agent finds a new path where there are no collisions with any of the already planned agents. This new path can then be used as the target planned path for the agent, and the three-dimensional time window can store the travel time of the agent at each path node along this target planned path.

[0061] The technical solution provided in this application, in response to a task scheduling instruction for any agent to be planned, firstly determines the initial planning path of the agent to be planned based on a pre-constructed road adjacency matrix, thereby determining a corresponding three-dimensional time window. This three-dimensional time window includes the travel time period when each agent passes through each path node in its corresponding planning path. Then, based on the first travel time period when the agent to be planned passes through each path node in the initial planning path and the second travel time period when the planned agents pass through each path node in the corresponding target planning path, the initial planning path is updated to obtain the target planning path for the agent to be planned. This ensures that there are no collisions between the agent to be planned and any of the planned agents in the target planning path, thereby achieving collision-free path planning for any agent in a multi-agent task scheduling scenario. This reduces resource overhead during multi-agent path planning, avoids deadlock in multi-agent task scheduling, improves the movement safety of any agent on the target planning path, and ensures the successful execution of the scheduling task by any agent.

[0062] As an optional implementation scheme in this application, considering that collision conflicts between the agent to be planned and any of the planned agents may exist when both pass through a certain path node at the same time, or when both travel towards each other in a road segment between two adjacent path nodes, this application can comprehensively analyze whether collision conflicts exist between the agent to be planned and the planned agents from different perspectives in order to completely avoid potential collision conflicts between them.

[0063] Next, this application will provide a detailed explanation of the specific update process for determining the target planning path of the agent to be planned by updating the initial planning path.

[0064] Figure 4 A flowchart illustrating another path planning method as shown in an embodiment of this application is as follows: Figure 4 As shown, the method may include the following steps:

[0065] S401, in response to the task scheduling instruction of the agent to be planned, determines the initial planning path of the agent to be planned based on the pre-constructed road adjacency matrix.

[0066] S402, Based on the initial planned path, determine the corresponding three-dimensional time window, which includes the time period when each agent passes through each path node in the planned path corresponding to the agent.

[0067] S403: Use the initial planned path as the current planned path, and use the first path node in the initial planned path as the current path node.

[0068] Considering that the agent to be planned may collide with a pre-planned agent while entering any path node or leaving any path node but not entering the next path node, this application needs to determine whether there is a collision conflict between the agent to be planned and the pre-planned agent at each path node in order to comprehensively avoid collision conflicts during the planning of the path by the agent to be planned.

[0069] Therefore, this application can use the initial planned path as the current planned path, and take the first path node in the initial planned path, which is the starting path node of the agent to be planned, as the current path node, so as to continuously update the initial planned path by traversing each path node.

[0070] S404, determine whether the first type of planned agent under the current path node is empty. If yes, execute S410; otherwise, execute S405.

[0071] For each current path node, the first step is to traverse all path nodes in the planned target path to check if the current path node exists in the target path. This allows us to identify the first type of planned agents that will pass through the current path node. In other words, the current path node exists in the target path of the first type of planned agents.

[0072] It should be understood that if the first type of planned agent under the current path node is empty, it means that there are no planned agents within the current path node or the road segment associated with the current path node, and therefore there is no collision conflict. Therefore, the next path node of the current path node can be directly used as the new current path node to continue determining the collision conflict between the planned agent and the planned agents.

[0073] For example, suppose the target planning path planned by agent A is 1→2→3→4, the target planning path planned by agent B is 4→3→6→2, and the initial planning path of agent C to be planned is 3→2→1→8.

[0074] Therefore, the first path node 3 is the current path node. The current path node 3 exists in the planned target paths of both planned agent A and planned agent B. Therefore, both planned agent A and planned agent B are first-class planned agents.

[0075] S405, based on the travel time of the agent to be planned and the first type of planned agent under the current path node, determine whether there is a first type collision between the agent to be planned and any first type of planned agent under the current path node. If yes, execute S406; if no, execute S407.

[0076] In this application, based on the location of the collision between the agent to be planned and the planned agent, the collision can be divided into two categories: one is the collision that occurs when the agent to be planned and a planned agent pass through the same path node at the same time, and the other is the collision that occurs when the agent to be planned and a planned agent travel towards each other in the same road segment formed between two adjacent path nodes at the same time.

[0077] Therefore, after identifying the first type of planned agents that have a current path node in the planned path, the second transit time of each first type of planned agent when passing through the current path node and the first transit time of the agent to be planned when passing through the current path node are first determined from the three-dimensional time window. Then, for any first type of planned agent, it is determined whether there is an overlap between the first transit time of the agent to be planned under the current path node and the second transit time of the first type of planned agent under the current path node.

[0078] If the time periods of the planned agent and a certain planned agent of type I overlap when passing through the current path node, it means that the two will pass through the current path node at the same time, and it can be determined that there is a type I collision between the planned agent and the planned agent of type I at the current path node.

[0079] In this context, a first-type collision refers to a collision that occurs when the agent to be planned and a pre-planned agent of the first type simultaneously pass through a path node. First-type collisions can be avoided by controlling the agent to wait for the pre-planned agent of the first type to exit the path node before entering it. Therefore, in this application, a first-type collision between the agent to be planned and any pre-planned agent of the first type at a path node is an avoidable collision.

[0080] For example, such as Figure 5 As shown, it is assumed that the first travel time of the agent to be planned under the current path node includes the first entry time point Q1 and the first exit time point Q2 under the current path node, and the second travel time of any first-class planned agent under the current path node includes the second entry time point P1 and the second exit time point P2 under the current path node.

[0081] Therefore, the condition for determining whether there is a first-type collision between the agent to be planned and any first-type planned agent at the current path node can be:

[0082] The above conditions can be interpreted as follows: the agent to be planned will exit the current path node after the first type of planned agent has entered it, and the agent to be planned will enter the current path node before the first type of planned agent has exited it. Therefore, the two agents must have overlapping travel periods at the current path node.

[0083] Therefore, if the entry and exit times of the agent to be planned and a certain first-type planned agent at the current path node meet the above conditions, it can be determined that there is a first-type collision between the agent to be planned and the first-type planned agent at the current path node.

[0084] S406, Update the first transit time of the agent to be planned at the current path node and each subsequent path node.

[0085] If a planned agent and a pre-planned agent of type I collide at the current path node, then considering that the travel time of the pre-planned agent of type I at the current path node is planned and cannot be changed, this application can avoid the first type of collision between the planned agent and the pre-planned agent of type I at the current path node by delaying the time when the planned agent enters the current path node.

[0086] Therefore, based on the second transit time of a planned agent of type I that has a type I collision with the agent to be planned at the current path node, the first transit time of the agent to be planned at the current path node can be set to follow the second transit time, thus obtaining the waiting time of the agent to be planned at the current path node. Furthermore, since the first transit time of the agent to be planned at the current path node changes, the first transit time of the agent to be planned at each subsequent path node will also change accordingly. Therefore, by controlling the first transit time of the agent to be planned at each subsequent path node to be delayed by the corresponding waiting time, the first transit time of the agent to be planned at each subsequent path node is updated.

[0087] As an optional implementation of this application, updating the first transit time of the planned intelligent agent under the current path node and each subsequent path node can specifically be as follows: determining the collision delay time of the planned intelligent agent under the current path node based on the first entry time point of the planned intelligent agent under the current path node and the second exit time point of the target first type of planned intelligent agent under the current path node; updating the first entry time point and the first exit time point of the planned intelligent agent under the current path node and each subsequent path node based on the collision delay time.

[0088] In other words, the first type of planned intelligent agent that has a first type of collision with the intelligent agent to be planned is the target first type of planned intelligent agent in this application. In order to ensure that the intelligent agent to be planned can enter the current path node only after the target first type of planned intelligent agent has left the current path node, and thus avoid the first type of collision between the two at the current path node, this application can calculate the difference between the first entry time point of the intelligent agent to be planned at the current path node and the second exit time point of the target first type of planned intelligent agent at the current path node, as the collision delay time that the intelligent agent to be planned needs to wait at the current path node.

[0089] For example, assuming the first entry time of the planned agent under the current path node is Q1, and the second exit time of the first type of planned agent under the current path node is P2, then the collision delay time of the planned agent under the current path node is delayTime = P2 - Q1.

[0090] Then, based on the collision delay time, the first entry time and the first exit time of the agent to be planned at the current path node and each subsequent path node can be delayed by the collision delay time, so as to obtain the new first entry time and the first exit time of the agent to be planned at the current path node and each subsequent path node, thereby updating the corresponding passage time information of the agent to be planned in the three-dimensional time window.

[0091] S407, determine whether the second type of planned agent under the current path node is empty. If yes, execute S410; otherwise, execute S408.

[0092] After determining whether a first-type collision exists between the agent to be planned and any first-type planned agent at the current path node, regardless of whether a first-type collision exists, it is necessary to further determine whether there is a second-type planned agent traveling towards the agent to be planned in the road segment between the agent leaving the current path node and entering the next path node. Specifically, the target planning path planned by the second-type planned agent contains a reverse sub-path from the current path node to the next path node.

[0093] In other words, by traversing each path node in the target planning path planned by each planned agent, it is possible to find a subset of planned agents whose target planning path contains a reverse sub-path that is opposite to the current sub-path from the current path node to the next path node. These are considered the second type of planned agents in this application. Since the target planning path of the second type of planned agents must contain a current path node, the second type of planned agents belongs to a subset of the first type of planned agents.

[0094] For example, suppose the target planning path planned by agent A is 1→2→3→4, the target planning path planned by agent B is 4→3→6→2, and the initial planning path of agent C to be planned is 3→2→1→8.

[0095] Therefore, the first path node 3 is the current path node. The current sub-path from the current path node to the next path node is 3→2, and the reverse sub-path is 2→3. So, by traversing the target planning paths of planned agents A and B, we can determine whether the reverse sub-path 2→3 exists. Since the target planning path of planned agent A contains the reverse sub-path 2→3, and the target planning path of planned agent B does not contain the reverse sub-path 2→3, planned agent A is a second type of planned agent.

[0096] It should be understood that if the planned agent has no second-type planned agents under the current path node, it means that there are no planned agents in the current sub-path from the current path node to the next path node, and therefore there is no collision conflict. Therefore, the next path node of the current path node can be directly used as the new current path node to continue to determine the collision conflict between the planned agent and the planned agents.

[0097] S408, based on the time periods of the planned agent and the second-class planned agents under the current path node and the next path node, determine whether there is a second-class collision between the planned agent and any second-class planned agent in the current sub-path from the current path node to the next path node. If yes, execute S409; if no, execute S410.

[0098] After identifying the second type of planned agent, the second type of planned agent travels from the next path node to the current path node, and travels towards the agent to be planned in the road segment between the current path node and the next path node.

[0099] At this point, the second transit time period for each planned agent of the second type can be determined from the three-dimensional time window when passing through the current path node and the next path node. Furthermore, the second entry time and the second exit time of the planned agent of the second type during the second transit time period under the current path node will be greater than the second entry time and the second exit time of the second transit time period under the next path node. Moreover, the first entry time and the first exit time of the agent to be planned during the first transit time period when passing through the current path node and the next path node can be determined.

[0100] Based on the aforementioned time points, the time period for the agent to be planned to travel from the current path node to the next path node within the current sub-path can be determined as the time period between the first entry time point under the current path node and the first entry time point under the next path node. Similarly, the time period for any second-type planned agent to travel from the next path node to the current path node within the reverse sub-path can be determined as the time period between the second exit time point under the next path node and the second exit time point under the current path node. Then, by determining whether there is an overlap between the time period for the agent to be planned to travel within the current sub-path and the time period for each second-type planned agent to travel within the reverse sub-path, it can be determined whether a second-type collision exists between the agent to be planned and any second-type planned agent within the current sub-path. If the time period of the agent to be planned when it travels through the current sub-path overlaps with that of a second-type planned agent when it travels through the reverse sub-path, it means that the two will pass through a certain point in the current sub-path at the same time. Therefore, it can be determined that there is a second-type collision between the agent to be planned and the second-type planned agent in the current sub-path from the current path node to the next path node.

[0101] The second type of collision refers to a collision that occurs when the agent to be planned and a pre-planned agent of the second type simultaneously travel towards each other on a road segment between two adjacent path nodes. This type of collision cannot be avoided by waiting for a period of time and is therefore unavoidable.

[0102] For example, such as Figure 6As shown, assume that the first travel time of the agent to be planned under the current path node includes the first entry time point P1 and the first exit time point P2 under the current path node, and the first travel time under the next path node includes the first entry time point N1 and the first exit time point N2 under the next path node. For any second-type planned agent, the second travel time under the next path node includes the second entry time point M1 and the second exit time point M2 under the next path node, and the second travel time under the current path node includes the second entry time point Q1 and the second exit time point Q2 under the current path node.

[0103] Therefore, the condition for determining whether there is a second-type collision between the agent to be planned and any second-type planned agent in the current sub-path from the current path node to the next path node can be:

[0104] The above conditions can be interpreted as follows: the planned intelligent agent will only leave the current path node after the planned intelligent agent enters it, and the planned intelligent agent will only enter the next path node after the planned intelligent agent leaves the next path node.

[0105] Therefore, if the entry and exit times of the agent to be planned and a certain second-type planned agent at the current path node and the next path node meet the above conditions, it can be determined that there is a second-type collision between the agent to be planned and the second-type planned agent in the current sub-path from the current path node to the next path node.

[0106] S409, Update the target path sequence from the current path node to the target path node in the current planned path to obtain the new current planned path and the new next path node, and return to execute S408.

[0107] When a second-type collision occurs between the agent to be planned and a second-type planned agent in the current sub-path from the current path node to the next path node, considering that the second-type collision is unavoidable, in order to ensure the safe passage of the agent to be planned during task scheduling, it is necessary to start from the current path node and search for a new path from the current path node to the target path node, which is the target path sequence in this application. Then, keeping the path sequence before the current path node in the current planned path unchanged, the new path searched from the current path node to the target path node is updated into the current planned path, thus obtaining the new current planned path and the new next path node, and returning to S408, based on the travel time of the agent to be planned and the second-type planned agent under the current path node and the new next path node, it is further determined whether there is a second-type collision between the agent to be planned and any second-type planned agent in the current sub-path from the current path node to the new next path node.

[0108] As an optional implementation of this application, updating the target path sequence from the current path node to the target path node in the current planned path to obtain a new current planned path can be specifically as follows: updating the path length between the current path node and the next path node in the road adjacency matrix to a pre-set impassable marker; searching for the shortest path from the current path node to the target path node based on the updated road adjacency matrix to obtain the corresponding target path sequence; merging the preceding path sequence and the target path sequence in the current planned path to obtain a new current planned path.

[0109] In other words, to avoid re-searching for the same path with a second-type collision between the current path node and the target path node, this application can first set the path from the current path node to the next path node as impassable for the agent to be planned. Therefore, in the road adjacency matrix, the path length between the current path node and the next path node can be modified to a pre-defined impassable marker, such as infinity, to indicate that the path from the current path node to the next path node is impassable.

[0110] Then, considering that the preceding path sequence in the current planned path has undergone collision detection and there are no collision conflicts, this application, based on the path length between every two path nodes in the updated road adjacency matrix, can use shortest path algorithms such as the Floyd algorithm or Dijkstra's algorithm to search for a new shortest path from the current path node to the target path node, which serves as the target path sequence between the current path node and the target path node. Furthermore, by merging the preceding path sequence and the target path sequence in the current planned path, a new current planned path can be obtained.

[0111] For example, suppose the target planning path planned by agent A is 1→2→3→4, the target planning path planned by agent B is 4→3→6→2, and the initial planning path of agent C to be planned is 3→2→1→8.

[0112] If the planned agent A and the agent C to be planned experience a second-type collision within the current sub-path 3→2 from the current path node to the next path node, then the path length of the current sub-path 3→2 can be modified to infinity in the road adjacency matrix. Then, the shortest path from the current path node 3 to the target path node 8 can be searched again, let's say 3→5→6→8. Since there is no preceding path sequence before the current path node 3, 3→5→6→8 can become the new planned path, and path node 5 can replace path node 2, becoming the new next path node of the current path node 3.

[0113] It should be understood that, in order to ensure the accuracy of the path planning for the agent to be planned, it is necessary to ensure the accuracy of the transit time for each path node in the three-dimensional time window. Therefore, after planning a new current planned path for the agent to be planned, it is necessary to redetermine the first entry time and the first exit time when the agent passes through each path node in the new current planned path, so as to update the first transit time corresponding to the agent in the three-dimensional time window.

[0114] In other words, based on the travel speed and size of the agent to be planned, as well as the path length between every two adjacent path nodes in the new current planning path, the first entry time and the second exit time of the agent to be planned when passing through each path node in the new current planning path can be calculated sequentially. This allows the first passage time period corresponding to each path node in the new current planning path to be updated in the three-dimensional time window.

[0115] S410: Take the latest next path node of the current path node as the new current path node, and return to execute S404.

[0116] When the first type of planned agents is empty, the second type of planned agents is empty, or there is no second type collision between the agent to be planned and all the second type of planned agents in the current sub-path, the latest next path node of the current path node can be taken as the new current path node, and then return to S404 to continue to determine whether there is a collision conflict between the agent to be planned and each of the planned agents under the new current path node.

[0117] S411, if the new current path node is empty, then the latest current planned path is taken as the target planned path for the agent to be planned.

[0118] The collision and conflict determination is completed for each path node within the planned path from the starting path node to the target path node until the new current path node is empty. This ensures that the agent to be planned will not collide with the already planned agents at any path node within the finally updated planned path. Therefore, this application can use the latest current planned path as the target planned path for the agent to be planned, thereby realizing collision-free path planning for any agent in a multi-agent task scheduling scenario.

[0119] The technical solution provided in this application, in response to a task scheduling instruction for any agent to be planned, firstly determines the initial planning path of the agent to be planned based on a pre-constructed road adjacency matrix, thereby determining a corresponding three-dimensional time window. This three-dimensional time window includes the travel time period when each agent passes through each path node in its corresponding planning path. Then, based on the first travel time period when the agent to be planned passes through each path node in the initial planning path and the second travel time period when the planned agents pass through each path node in the corresponding target planning path, the initial planning path is updated to obtain the target planning path for the agent to be planned. This ensures that there are no collisions between the agent to be planned and any of the planned agents in the target planning path, thereby achieving collision-free path planning for any agent in a multi-agent task scheduling scenario. This reduces resource overhead during multi-agent path planning, avoids deadlock in multi-agent task scheduling, improves the movement safety of any agent on the target planning path, and ensures the successful execution of the scheduling task by any agent.

[0120] Figure 7 This is a schematic block diagram illustrating a path planning device according to an embodiment of this application. Figure 7 As shown, the device 700 may include:

[0121] The initial path determination module 710 is used to determine the initial planning path of the intelligent agent to be planned based on a pre-constructed road adjacency matrix in response to the task scheduling instruction of the intelligent agent to be planned.

[0122] The time window determination module 720 is used to determine the corresponding three-dimensional time window based on the initial planned path. The three-dimensional time window includes the time period when each agent passes through each path node in the planned path corresponding to the agent.

[0123] The path planning module 730 is used to update the initial planning path based on the first travel time when the agent to be planned passes through each path node in the initial planning path and the second travel time when the planned agent passes through each path node in the corresponding target planning path, so as to obtain the target planning path of the agent to be planned.

[0124] In some implementations, the path planning module 730 can be specifically used for:

[0125] The initial planned path is taken as the current planned path, and the first path node in the initial planned path is taken as the current path node;

[0126] Execution path update step: When it is determined that there is a first type of collision between the agent to be planned and any first type of planned agent under the current path node based on the travel time of the agent to be planned and the first type of planned agent under the current path node, update the first travel time of the agent to be planned under the current path node and each subsequent path node.

[0127] When it is determined, based on the travel time of the agent to be planned and the second type of planned agent under the current path node and the next path node, that there is a second type of collision between the agent to be planned and any second type of planned agent in the current sub-path from the current path node to the next path node, the target path sequence from the current path node to the target path node in the current planning path is updated to obtain a new current planning path and a new next path node, so as to continue to determine whether there is a second type of collision between the agent to be planned and any second type of planned agent in the current sub-path from the current path node to the new next path node;

[0128] If the first type of planned agent is empty, the second type of planned agent is empty, or the agent to be planned does not have any second-type collision with any of the second-type planned agents in the current sub-path, then the latest next path node of the current path node is taken as the new current path node, and the above path update steps are continued until the new current path node is empty, then the latest current planned path is taken as the target planned path of the agent to be planned.

[0129] In the first type of planned intelligent agent, the target planning path contains a current path node, and in the second type of planned intelligent agent, the target planning path contains a reverse sub-path from the current path node to the next path node.

[0130] In some implementations, the path planning module 730 can be specifically used for:

[0131] Based on the first entry time of the planned intelligent agent under the current path node and the second exit time of the first type of planned intelligent agent under the current path node, the collision delay time of the planned intelligent agent under the current path node is determined.

[0132] Based on the collision delay duration, update the first entry time and first exit time of the agent to be planned at the current path node and each subsequent path node.

[0133] In some implementations, the path planning module 730 can also be specifically used for:

[0134] Update the path length between the current path node and the next path node in the road adjacency matrix to a preset impassable flag;

[0135] Based on the updated road adjacency matrix, search for the shortest path from the current path node to the target path node to obtain the corresponding target path sequence;

[0136] The preceding path sequence and the target path sequence located before the current path node in the current planned path are merged to obtain a new current planned path.

[0137] In some implementations, the path planning device 700 may further include:

[0138] The time window update module is used to determine the first entry time and the first exit time when the intelligent agent to be planned passes through each path node in the new current planned path, and update the first passage time period corresponding to the intelligent agent to be planned in the three-dimensional time window.

[0139] In some implementations, the initial path determination module 710 can be specifically used for:

[0140] Based on the path nodes in the road scene map where the agent to be planned is located, a corresponding road adjacency matrix is ​​constructed, and the road adjacency matrix includes the path length between every two path nodes;

[0141] Based on the road adjacency matrix, the shortest path between the starting path node and the target path node of the agent to be planned is searched to obtain the initial planned path of the agent to be planned.

[0142] In this embodiment, in response to a task scheduling instruction for any agent to be planned, the initial planning path of the agent is first determined based on a pre-constructed road adjacency matrix to determine a corresponding three-dimensional time window. This three-dimensional time window includes the travel time when each agent passes through each path node in its corresponding planning path. Then, based on the first travel time when the agent to be planned passes through each path node in the initial planning path and the second travel time when the planned agents pass through each path node in the corresponding target planning path, the initial planning path is updated to obtain the target planning path for the agent to be planned. This ensures that there are no collisions between the agent to be planned and any of the planned agents in the target planning path, thereby achieving collision-free path planning for any agent in a multi-agent task scheduling scenario. This reduces resource overhead during multi-agent path planning, avoids deadlock in multi-agent task scheduling, improves the movement safety of any agent on the target planning path, and ensures the successful execution of the scheduling task by any agent.

[0143] It should be understood that the device embodiments and method embodiments can correspond to each other, and similar descriptions can be referred to the method embodiments. To avoid repetition, further details will not be provided here. Specifically, Figure 7 The apparatus 700 shown can execute any of the method embodiments in this application, and the foregoing and other operations and / or functions of each module in the apparatus 700 are respectively for implementing the corresponding processes in the various methods in the embodiments of this application. For the sake of brevity, they will not be described in detail here.

[0144] The apparatus 700 of this application embodiment has been described above from the perspective of functional modules in conjunction with the accompanying drawings. It should be understood that this functional module can be implemented in hardware, in software instructions, or in a combination of hardware and software modules. Specifically, the steps of the method embodiments in this application can be completed by integrated logic circuits in the processor's hardware and / or by software instructions. The steps of the method disclosed in this application embodiment can be directly embodied as being executed by a hardware decoding processor, or by a combination of hardware and software modules in the decoding processor. Optionally, the software module can reside in a mature storage medium in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, etc. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps in the above method embodiments.

[0145] Figure 8 This is a schematic block diagram of an electronic device shown in an embodiment of this application.

[0146] like Figure 8 As shown, the electronic device 800 may include:

[0147] The system includes a memory 810 and a processor 820. The memory 810 stores computer programs and transfers the program code to the processor 820. In other words, the processor 820 can retrieve and run the computer program from the memory 810 to implement the methods described in the embodiments of this application.

[0148] For example, the processor 820 can be used to execute the above-described method embodiments according to instructions in the computer program.

[0149] In some embodiments of this application, the processor 820 may include, but is not limited to:

[0150] General-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0151] In some embodiments of this application, the memory 810 includes, but is not limited to:

[0152] Volatile memory and / or non-volatile memory. Non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), and Direct Rambus RAM (DR RAM).

[0153] In some embodiments of this application, the computer program may be divided into one or more modules, which are stored in the memory 810 and executed by the processor 820 to perform the method provided in this application. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the electronic device.

[0154] like Figure 8 As shown, the electronic device may further include:

[0155] Transceiver 830, which can be connected to processor 820 or memory 810.

[0156] The processor 820 can control the transceiver 830 to communicate with other devices; specifically, it can send information or data to other devices or receive information or data sent by other devices. The transceiver 830 may include a transmitter and a receiver. The transceiver 830 may further include antennas, and the number of antennas may be one or more.

[0157] It should be understood that the various components in the electronic device are connected through a bus system, which includes a data bus, a power bus, a control bus, and a status signal bus.

[0158] This application also provides a computer storage medium storing a computer program thereon, which, when executed by a computer, enables the computer to perform the methods of the above-described method embodiments. Alternatively, embodiments of this application also provide a computer program product containing instructions that, when executed by a computer, cause the computer to perform the methods of the above-described method embodiments.

[0159] When implemented using software, it can be implemented entirely or partially as a computer program product. This computer program product includes one or more computer instructions. When these computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., digital video disc (DVD)), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0160] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0161] In the several embodiments provided in this application, 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 modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules 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 apparatuses or modules may be electrical, mechanical, or other forms.

[0162] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. For example, the functional modules in the various embodiments of this application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0163] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A path planning method characterized by, include: In response to the task scheduling instruction of the agent to be planned, the initial planning path of the agent to be planned is determined according to the pre-constructed road adjacency matrix; Based on the initial planned path, a corresponding three-dimensional time window is determined. The three-dimensional time window includes the time period when each agent passes through each path node in the planned path corresponding to the agent. The initial planning path is updated based on the first travel time when the agent to be planned passes through each path node in the initial planning path and the second travel time when the planned agent passes through each path node in the corresponding target planning path, so as to obtain the target planning path of the agent to be planned. The step of updating the initial planning path based on the first travel time when the agent to be planned passes through each path node in the initial planning path and the second travel time when the planned agent passes through each path node in the corresponding target planning path, to obtain the target planning path for the agent to be planned, includes: The initial planned path is taken as the current planned path, and the first path node in the initial planned path is taken as the current path node; Execution path update step: When it is determined that there is a first type of collision between the agent to be planned and any first type of planned agent under the current path node based on the travel time of the agent to be planned and the first type of planned agent under the current path node, update the first travel time of the agent to be planned under the current path node and each subsequent path node. When it is determined, based on the travel time of the agent to be planned and the second type of planned agent under the current path node and the next path node, that there is a second type of collision between the agent to be planned and any second type of planned agent in the current sub-path from the current path node to the next path node, the target path sequence from the current path node to the target path node in the current planning path is updated to obtain a new current planning path and a new next path node, so as to continue to determine whether there is a second type of collision between the agent to be planned and any second type of planned agent in the current sub-path from the current path node to the new next path node; If the first type of planned agent is empty, the second type of planned agent is empty, or the agent to be planned does not have any second-type collision with any of the second-type planned agents in the current sub-path, then the latest next path node of the current path node is taken as the new current path node, and the above path update steps are continued until the new current path node is empty, then the latest current planned path is taken as the target planned path of the agent to be planned. In the first type of planned intelligent agent, the target planning path contains a current path node, and in the second type of planned intelligent agent, the target planning path contains a reverse sub-path from the current path node to the next path node.

2. The method according to claim 1, characterized in that, The step of updating the first transit time period of the agent to be planned at the current path node and each subsequent path node includes: Based on the first entry time of the planned intelligent agent under the current path node and the second exit time of the first type of planned intelligent agent under the current path node, the collision delay time of the planned intelligent agent under the current path node is determined. Based on the collision delay duration, update the first entry time and first exit time of the agent to be planned at the current path node and each subsequent path node.

3. The method according to claim 1, characterized in that, The step of updating the target path sequence from the current path node to the target path node in the current planned path to obtain a new current planned path includes: Update the path length between the current path node and the next path node in the road adjacency matrix to a preset impassable flag; Based on the updated road adjacency matrix, search for the shortest path from the current path node to the target path node to obtain the corresponding target path sequence; The preceding path sequence and the target path sequence located before the current path node in the current planned path are merged to obtain a new current planned path.

4. The method according to claim 1, characterized in that, After updating the target path sequence from the current path node to the target path node in the current planned path to obtain a new current planned path, the method further includes: Determine the first entry time and the first exit time when the agent to be planned passes through each path node in the new currently planned path, and update the first passage time period corresponding to the agent to be planned in the three-dimensional time window.

5. The method according to claim 1, characterized in that, The step of determining the initial planning path of the agent to be planned based on the pre-constructed road adjacency matrix includes: Based on the path nodes in the road scene map where the agent to be planned is located, a corresponding road adjacency matrix is ​​constructed, and the road adjacency matrix includes the path length between every two path nodes; Based on the road adjacency matrix, the shortest path between the starting path node and the target path node of the agent to be planned is searched to obtain the initial planned path of the agent to be planned.

6. A path planning device, characterized in that, include: The initial path determination module is used to determine the initial planning path of the agent to be planned based on a pre-constructed road adjacency matrix in response to the task scheduling instruction of the agent to be planned. The time window determination module is used to determine the corresponding three-dimensional time window based on the initial planned path. The three-dimensional time window includes the time period when each agent passes through each path node in the planned path corresponding to the agent. The path planning module is used to update the initial planning path based on the first travel time when the agent to be planned passes through each path node in the initial planning path and the second travel time when the planned agent passes through each path node in the corresponding target planning path, so as to obtain the target planning path of the agent to be planned. The path planning module is specifically used for: The initial planned path is taken as the current planned path, and the first path node in the initial planned path is taken as the current path node; Execution path update step: When it is determined that there is a first type of collision between the agent to be planned and any first type of planned agent under the current path node based on the travel time of the agent to be planned and the first type of planned agent under the current path node, update the first travel time of the agent to be planned under the current path node and each subsequent path node. When it is determined, based on the travel time of the agent to be planned and the second type of planned agent under the current path node and the next path node, that there is a second type of collision between the agent to be planned and any second type of planned agent in the current sub-path from the current path node to the next path node, the target path sequence from the current path node to the target path node in the current planning path is updated to obtain a new current planning path and a new next path node, so as to continue to determine whether there is a second type of collision between the agent to be planned and any second type of planned agent in the current sub-path from the current path node to the new next path node; If the first type of planned agent is empty, the second type of planned agent is empty, or the agent to be planned does not have any second-type collision with any of the second-type planned agents in the current sub-path, then the latest next path node of the current path node is taken as the new current path node, and the above path update steps are continued until the new current path node is empty, then the latest current planned path is taken as the target planned path of the agent to be planned. In the first type of planned intelligent agent, the target planning path contains a current path node, and in the second type of planned intelligent agent, the target planning path contains a reverse sub-path from the current path node to the next path node.

7. An electronic device, characterized in that, include: A processor and a memory, the memory being used to store a computer program, the processor being used to invoke and run the computer program stored in the memory to perform the path planning method according to any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, Used to store a computer program that causes a computer to perform the path planning method as described in any one of claims 1-5.

9. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instruction is executed by the processor, it implements the path planning method as described in any one of claims 1-5.

Images