Robot path planning method, device, system and readable storage medium

Abstract

The application discloses a path planning method, device and system of a robot, and a readable storage medium. The path planning method comprises the following steps: performing path planning on a first robot and a second robot to obtain a first planning path of the first robot and a second planning path of the second robot; performing head-on motion conflict detection on the first robot and the second robot according to the first planning path and the second planning path; the head-on motion conflict detection comprises robot size conflict detection; in response to the existence of head-on motion conflict between the first robot and the second robot, seeking an avoidance node for the first robot or the second robot, and performing action constraint on the first robot or the second robot; and performing path planning on the first robot and the second robot again according to the avoidance node and the action constraint until the final planning path of the first robot and the second robot is obtained. In this way, the problem of continuous conflict resolution in a space-limited scene can be reduced.

Description

Technical Field

[0001] This application relates to the field of robot path planning technology, and in particular to robot path planning methods, apparatus, systems, and readable storage media. Background Technology

[0002] Multi-robot systems are increasingly used in flexible manufacturing and smart manufacturing, but the diversification and complexity of application scenarios also pose challenges to the robustness and efficient operation of these systems. Multi-robot path planning (MRPP) is key to ensuring efficient collaborative operation of robots. MRPP aims to plan a conflict-free path for each robot on the map, enabling each robot to move from its starting position to its target position.

[0003] The Conflict-Based Search (CBS) algorithm is a two-layer search method. The upper layer is responsible for detecting conflicts for the robot and constructing a constraint tree (CT) based on the conflict time and location. The lower layer uses a single-path solver to plan the shortest path for each robot that satisfies the constraints in the CT nodes.

[0004] However, the conflict search algorithm, which relies on node location comparison, may miss conflicts between opposing lanes, leading to the continuous resolution of conflicts and reducing the solution efficiency. Summary of the Invention

[0005] This application provides a path planning method, apparatus, system, and readable storage medium for robots, which can reduce the persistent problem of conflict resolution in space-constrained scenarios.

[0006] To solve the above-mentioned technical problems, one technical solution adopted in this application is: providing a path planning method for a robot, the robot including at least a first robot and a second robot, the path planning method including: performing path planning on the first robot and the second robot to obtain a first planned path for the first robot and a second planned path for the second robot; performing mutual motion conflict detection on the first robot and the second robot based on the first planned path and the second planned path; mutual motion conflict detection including robot size conflict detection during mutual motion; in response to the existence of mutual motion conflict between the first robot and the second robot, finding an avoidance node for the first robot or the second robot, and imposing motion constraints on the first robot or the second robot; re-planning the path for the first robot and the second robot based on the avoidance node and motion constraints, until the final planned path for the first robot and the second robot is obtained.

[0007] The process of finding an avoidance node for the first robot or the second robot in response to a conflict of movement between them, and imposing action constraints on the first robot or the second robot, includes: obtaining the conflict time corresponding to the conflict of movement between the first robot and the second robot in response to a conflict of movement between them; finding an avoidance node for the first robot or the second robot based on the conflict time; and imposing action constraints on the first robot or the second robot.

[0008] The process of finding an avoidance node for the first or second robot based on the moment of conflict and imposing action constraints on the first or second robot includes: backtracking the first and second planned paths step by step based on the moment of conflict; and finding an avoidance node based on node position and path information during the step-by-step backtracking process and imposing action constraints on the first or second robot.

[0009] Among them, the opposing motion conflict includes robot size conflict; during the time-step backtracking process, avoidance nodes are found based on node positions and path information, and action constraints are imposed on the first robot, including: using the conflict time as a reference, performing robot size conflict detection on the first robot in descending order of time steps and on the second robot in ascending order of time steps; and using the conflict time as a reference, performing robot size conflict detection on the first robot in ascending order of time steps and on the second robot in descending order of time steps to obtain a size conflict index table corresponding to the first robot and the second robot; in response to the first robot needing to avoid the second robot, a waiting action is generated for the second robot at the node corresponding to the previous time step at the conflict time, and the... The robot performs size conflict detection on all actions and waiting actions of the first robot in the previous moment, generating action constraints for the first robot. Based on the size conflict index table, the robot size conflict detection is performed on the occupant model of the first robot at the conflict moment and the occupant model of the second robot at all moments, determining the conflict termination moment. Based on the earliest moment when the first robot arrives at the node where the conflict moment is located and the conflict termination moment, all action constraints of the first robot relative to the node where the conflict moment is located are generated. Based on the size conflict index table, the time steps are traversed in reverse with the conflict moment as the reference, and preset operations are performed on the target node where the first robot is located at each moment to find avoidance nodes and impose action constraints on the first robot.

[0010] Specifically, based on the size conflict index table, the time steps are traversed backwards from the conflict time as a reference. Preset operations are performed on the target node where the first robot is located at each time step to find avoidance nodes and impose motion constraints on the first robot. This includes: based on the size conflict index table, the time steps are traversed backwards from the conflict time as a reference, and the following operations are performed on the target node where the first robot is located at each time step: for each target node, a set of nodes connected to the target node at each time step is obtained; and robot size conflict detection is performed on the occupancy model of the first robot at the target node and the occupancy models of the second robot at all times, determining the potential conflict termination time and recording it. Record the earliest time when the first robot arrives at the target node; detect whether there is a conflict between the occupant model corresponding to the node in the node set from the target node and the occupant model of the second robot in the first time interval; the first time interval is the time between the moment before the conflict moment and the conflict moment; if so, determine the second time interval based on the earliest time when the first robot arrives at the target node and the potential conflict termination moment, and impose action constraints on the first robot in the second time interval; if not, determine that there is an avoidance node in the node set, determine the third time interval based on the earliest time when the first robot arrives at the target node and the potential conflict termination moment, and impose action constraints on the first robot in the third time interval.

[0011] Specifically, a second time interval is determined based on the earliest time when the first robot arrives at the target node and the potential conflict termination time, and the actions of the first robot are constrained within the second time interval, including: defining the period from the earliest time when the first robot arrives at the target node to the first potential conflict termination time as the second time interval; the first potential conflict termination time being later than the potential conflict termination time; and constraining all actions of the first robot at the target node within the second time interval.

[0012] Specifically, a third time interval is determined based on the earliest time when the first robot arrives at the target node and the time when the potential conflict ends. Within the third time interval, the actions of the first robot are constrained, including: defining the time interval from the target time when the first robot is at the target node to the time when the first potential conflict ends; the target time is later than the earliest time when the first robot arrives at the target node, and the time when the first potential conflict ends is later than the time when the potential conflict ends; within the third time interval, the waiting actions and moving actions of the first robot at the target node are constrained, while the rotation action is retained.

[0013] To solve the above-mentioned technical problems, another technical solution adopted in this application is to provide a path planning device for a robot, which includes a processor, a memory, and a communication module. The memory is used to store computer programs, and the processor is used to execute the computer programs to implement the method provided by the above-mentioned technical solution.

[0014] To solve the above-mentioned technical problems, another technical solution adopted in this application is to provide a robot system, which includes: multiple robots, including at least a first robot and a second robot; and a robot path planning device, such as the robot path planning device provided in the above-mentioned technical solution.

[0015] To solve the above-mentioned technical problems, another technical solution adopted in this application is to provide a computer-readable storage medium that stores a computer program, which, when executed by a processor, is used to implement the method provided by the above-mentioned technical solution.

[0016] The robot path planning method, apparatus, system, and readable storage medium provided in this application, after performing path planning on a first robot and a second robot to obtain a first planned path for the first robot and a second planned path for the second robot, performs mutual motion conflict detection on the first robot and the second robot based on the first planned path and the second planned path. For example, mutual motion conflict detection includes robot size conflict detection during mutual motion. In this way, when there is mutual motion conflict between the first robot and the second robot, an avoidance node is found for the first robot or the second robot, and motion constraints are applied to the first robot or the second robot. Based on the avoidance node and motion constraints, the path planning of the first robot and the second robot is re-performed until the final planned path of the first robot and the second robot is obtained. It can effectively identify size mutual conflict in narrow space resource-constrained scenarios such as alleyways, find avoidance paths for the conflicting first robot and the second robot, reduce the number of times such conflicts are resolved, and improve solution efficiency. Attached Figure Description

[0017] 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 schematic diagram of the structure of an embodiment of the robot system provided in this application; Figure 2 This is a flowchart illustrating an embodiment of the robot path planning method provided in this application; Figure 3 This is a flowchart illustrating another embodiment of the robot path planning method provided in this application; Figure 4 yes Figure 3 A flowchart illustrating an embodiment of step 34; Figure 5 yes Figure 4 A flowchart illustrating an embodiment of step 342; Figure 6 yes Figure 5 A flowchart of an embodiment of step 55; Figure 7 This is a schematic diagram of the roadway conflict between AGV1 and AGV2 provided in this application; Figure 8 This is a schematic diagram of the conflict and avoidance of the roadway dimensions of AGV1 and AGV2 provided in this application; Figure 9 This is a schematic diagram of an embodiment of the robot path planning device provided in this application; Figure 10 This is a schematic diagram of an embodiment of the computer-readable storage medium provided in this application. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It is understood that the specific embodiments described herein are only for explaining this application and not for limiting it. Furthermore, it should be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all structures. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0019] In this document, the term "embodiment" means that a particular 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.

[0020] Multi-robot systems are increasingly used in flexible manufacturing and smart manufacturing, but the diversification and complexity of application scenarios also pose challenges to the robustness and efficient operation of these systems. Multi-robot path planning (MRPP) is key to ensuring efficient collaborative operation of robots. MRPP aims to plan a conflict-free path for each robot on the map, enabling each robot to move from its starting position to its target position.

[0021] The Conflict-Based Search (CBS) algorithm is a two-layer search method. The upper layer is responsible for detecting conflicts for the robot and constructing a constraint tree (CT) based on the conflict time and location. The lower layer uses a single-path solver to plan the shortest path for each robot that satisfies the constraints in the CT nodes.

[0022] However, the conflict search algorithm, which relies on node location comparison, may miss conflicts between opposing lanes, leading to the continuous resolution of conflicts and reducing the solution efficiency.

[0023] Based on this, this application proposes that after path planning is performed on the first robot and the second robot to obtain the first planned path of the first robot and the second planned path of the second robot, a conflict detection of opposing motion is performed on the first robot and the second robot according to the first planned path and the second planned path. For example, the conflict detection of opposing motion includes the detection of robot size conflict during opposing motion. In this way, when there is a conflict of opposing motion between the first robot and the second robot, an avoidance node is found for the first robot or the second robot, and motion constraints are applied to the first robot or the second robot. Based on the avoidance node and motion constraints, the path planning of the first robot and the second robot is re-performed until the final planned path of the first robot and the second robot is obtained. This can effectively identify the size conflict of opposing motion in narrow spaces such as alleys, find an avoidance path for the conflicting first robot and the second robot, reduce the number of times such conflicts are resolved, and improve the solution efficiency. See any of the following embodiments for details.

[0024] See Figure 1 , Figure 1 This is a schematic diagram of an embodiment of the robot system provided in this application. The robot system 100 includes: a robot path planning device 30 and multiple robots. The multiple robots include at least a first robot 10 and a second robot 20. The robot path planning device 30 can perform corresponding path planning for the multiple robots, so as to plan a non-conflicting path for each robot in the corresponding scenario. The robots can be handling robots, such as handling robots applied in logistics scenarios. See any of the following embodiments for details.

[0025] See Figure 2 , Figure 2 This is a flowchart illustrating an embodiment of the robot path planning method provided in this application. The robot includes at least a first robot and a second robot, and the path planning method includes: Step 21: Perform path planning for the first robot and the second robot to obtain the first planned path for the first robot and the second planned path for the second robot.

[0026] In some embodiments, each path consists of several nodes. That is, a node can refer to the smallest unit of a map, and is also the smallest search unit for path planning.

[0027] In some embodiments, path planning can be performed on the first robot and the second robot based on the map of the current scene, the current positions of the first robot and the second robot, and the target positions corresponding to the tasks performed by the first robot and the second robot, to obtain a first planned path for the first robot and a second planned path for the second robot.

[0028] In some embodiments, the robot includes at least a first robot, a second robot, and a third robot, and path planning can be performed on the first robot, the second robot, and the third robot to obtain the planned paths corresponding to the first robot, the second robot, and the third robot, respectively.

[0029] Step 22: Based on the first and second planned paths, perform collision detection on the first robot and the second robot moving towards each other.

[0030] In some embodiments, in addition to performing the opposing motion conflict detection in step 22, point conflict detection, edge conflict detection, following conflict detection, and size conflict detection also need to be performed.

[0031] Point collision detection: For robot a i And robot a j path p i and p j If their actions to move to their respective target node positions at time t conflict, that is, if there is a vertex conflict.

[0032] Edge collision detection: For robot a i And robot a j path p i and p j If their movements conflict between time t-1 and time t, i.e., robot a... i And robot a j An edge collision occurs at time t.

[0033] Follow-up collision detection: If robot a i path p i The intention is to reach node v at time t, while robot a... j path p j The action at the moment before t will be related to robot a. i The action of moving to node v is conflicting, i.e., robot a i And robot a j A following conflict occurs at time t.

[0034] Conflict information is generated based on collision comparison of moving lock grids. i ;a j action ai action aj ;t> represents robot a i And robot a j At time t, its action information is action. ai and action aj A conflict occurred. ai Robot a i Action information, action aj Robot a j Action information.

[0035] In some embodiments, a two-layer search architecture of CBS can be adopted: the higher layer performs conflict search and constructs a CT tree, where CT nodes store robot path sets, path costs, multi-machine conflict information, and corresponding constraints; the lower layer single-machine path solver generates the optimal path for the robot based on the CT node constraints. Additionally, the node in the CT tree's open table that satisfies the minimum cost is multiplied by a suboptimal coefficient and stored in the focal table for CT node selection. A conflict detection mechanism based on kinematic modeling (opposite motion conflict detection) is used: a motion lock grid is generated step-by-step for conflict determination, and a constraint is added to one of the robots prohibiting it from performing actions that conflict with another robot at conflict time t.

[0036] The mobile grid representation is based on the robot body and load model, combined with kinematic actions such as waiting, moving forward, moving backward, arc motion, and rotation, to construct a two-dimensional spatial occupancy model of the robot and load combination in the map.

[0037] In some embodiments, the collision detection for opposing motion includes collision detection during opposing motion. That is, when robots move towards each other, in addition to detecting collisions at node positions, collisions may still occur even if the node positions do not conflict, due to differences in robot size and motion. Therefore, collision detection for robot size conflicts is necessary.

[0038] For example, to verify whether the following two sets of timing sequences of moving lock cells simultaneously have secondary conflicts at the conflict time T: (1) The movement lock of AGV1 at time T-1 and the movement lock of AGV2 at time T+1.

[0039] (2) The movement and locking of AGV2 at time T-1 and the movement and locking of AGV1 at time T+1.

[0040] ​When a conflicting node is located at a path endpoint (starting or ending point), if any one of the two sets of moving lock cells satisfying the above two time sequences exists, it can be determined as a roadway size conflict, and the conflict type is updated. This detection process can be understood as a supplement to time-step moving lock cell detection. Considering that moving lock cell detection takes longer than node position comparison, if the conflict can be determined to be a roadway conflict through node position comparison, then roadway size conflict detection is unnecessary. Traditional roadway conflict detection methods based on node position matching cannot determine this type of conflict. Since there is a moving lock cell conflict before and after the conflict time T, it is determined to be a roadway size opposing conflict.

[0041] Step 23: In response to the conflict of movement between the first robot and the second robot, find an avoidance node for the first robot or the second robot, and impose motion constraints on the first robot or the second robot.

[0042] In some embodiments, since there is a conflict of movement between the first robot and the second robot, an avoidance node can be found for the first robot, and the action constraints of the first robot can be applied.

[0043] In some embodiments, since there is a conflict of movement between the first robot and the second robot, an avoidance node can be found for the second robot, and the action constraints of the second robot can be applied.

[0044] Step 24: Based on the avoidance nodes and motion constraints, re-plan the paths for the first and second robots until the final planned paths for the first and second robots are obtained.

[0045] In some embodiments, the final planned paths of the first robot and the second robot ensure that their movements do not conflict with each other.

[0046] In this embodiment, after path planning is performed on the first robot and the second robot to obtain the first planned path of the first robot and the second planned path of the second robot, the first robot and the second robot are subjected to mutual motion conflict detection based on the first planned path and the second planned path. For example, mutual motion conflict detection includes robot size conflict detection when moving towards each other. In this way, when there is mutual motion conflict between the first robot and the second robot, an avoidance node is found for the first robot or the second robot, and motion constraints are applied to the first robot or the second robot. Based on the avoidance node and motion constraints, the first robot and the second robot are re-planned until the final planned path of the first robot and the second robot is obtained. This can effectively identify size conflict in narrow passage scenarios such as alleys, find avoidance paths for the conflicting first robot and the second robot, reduce the number of times such conflicts are resolved, and improve the solution efficiency.

[0047] See Figure 3 , Figure 3 This is a flowchart illustrating another embodiment of the robot path planning method provided in this application. The robot includes at least a first robot and a second robot, and the path planning method includes: Step 31: Perform path planning for the first robot and the second robot to obtain the first planned path for the first robot and the second planned path for the second robot.

[0048] Step 32: Based on the first and second planned paths, perform collision detection on the first robot and the second robot moving towards each other.

[0049] In some embodiments, steps 31 to 32 have the same or similar technical solutions as any embodiment of this application, and will not be described in detail here.

[0050] Step 33: In response to the conflict between the first robot and the second robot, obtain the conflict time corresponding to the conflict.

[0051] In some embodiments, by performing conflict detection based on the first and second planned paths, the moment of conflict corresponding to the conflict can be determined. Based on this, the nodes where the first and second robots are located at the moment of conflict can be clearly known.

[0052] Step 34: Find an avoidance point for the first robot or the second robot based on the time of conflict, and impose motion constraints on the first robot or the second robot.

[0053] In some embodiments, avoidance nodes can be found for the first robot, and action constraints can be imposed on the first robot.

[0054] In some embodiments, avoidance nodes can be found for the second robot, and action constraints can be imposed on the second robot.

[0055] In some embodiments, see Figure 4 Step 34 can be the following process: Step 341: Using the moment of conflict as a reference, backtrack the first and second planning paths step by step.

[0056] In some embodiments, the first robot can be sorted in descending order by time step and the second robot in ascending order by time step, based on the moment of conflict, so as to backtrack the corresponding nodes in the first and second planned paths.

[0057] In some embodiments, the first robot can be sorted in ascending order by time step and the second robot in descending order by time step, based on the moment of conflict, so as to backtrack the corresponding nodes in the first and second planned paths.

[0058] Step 342: During the time-step backtracking process, avoidance nodes are found based on node positions and path information, and action constraints are imposed on the first robot or the second robot.

[0059] In some embodiments, during the time-step backtracking process, avoidance nodes are found for the first robot based on node positions and path information, and action constraints are imposed on the first robot.

[0060] In some embodiments, during the time-step backtracking process, avoidance nodes are found for the second robot based on node positions and path information, and action constraints are imposed on the second robot.

[0061] In some embodiments, opposing motion conflicts include robot size conflicts. See also Figure 5 Step 342 can be the following process: Step 51: Based on the conflict time, perform robot size conflict detection on the first robot in descending order of time steps and on the second robot in ascending order of time steps; and based on the conflict time, perform robot size conflict detection on the first robot in ascending order of time steps and on the second robot in descending order of time steps to obtain the size conflict index table corresponding to the first robot and the second robot.

[0062] In some embodiments, robot size conflict detection is performed on the first robot in descending time step order, such as t=T, T-1, T-2…, and on the second robot in ascending time step order, such as t=T, T+1, T+2…. T represents the conflict time.

[0063] In some embodiments, robot size conflict detection is performed on the first robot in ascending time step order, such as t=T, T+1, T+2… and on the second robot in descending time step order, such as t=T, T-1, T-2…

[0064] Based on this, during the robot size conflict detection process, when no new conflicts are generated, a size conflict index table corresponding to the first and second robots is obtained. The size conflict index table can contain relevant conflict information for each robot, such as the time of the conflict and the location of the conflict node.

[0065] Step 52: In response to the first robot needing to avoid the second robot, generate a waiting action for the second robot at the node corresponding to the previous moment at the time of conflict, and perform robot size conflict detection on all actions of the first robot at the previous moment and the waiting action to generate action constraints for the first robot.

[0066] In some embodiments, when the first robot is required to avoid the second robot, a waiting action is generated for the second robot at the node position at time T-1. This waiting action is then compared with all possible actions of the first robot at time T-1 using a lock-by-lock conflict detection method (robot size conflict detection). Detected conflicts are converted into constraints and added to the constraint set corresponding to the first robot, and used for subsequent CT child node inheritance. By pre-detecting the action of the node before the conflict occurs, a diagonal avoidance path is generated to eliminate unilateral avoidance size conflicts.

[0067] In some embodiments, when the second robot needs to avoid the first robot, a waiting action can be generated for the first robot at the node corresponding to the previous moment at the time of conflict, and robot size conflict detection can be performed on all actions of the second robot at the previous moment and the waiting action to generate action constraints for the second robot.

[0068] In some embodiments, since following conflicts need to be considered when resolving conflicts, it is necessary to avoid a situation where continuous conflict resolution occurs at the moment of conflict. Based on this, step 53 is performed.

[0069] In some embodiments, the size conflict index table may be a lane size conflict index table.

[0070] Step 53: Based on the size conflict index table, perform robot size conflict detection between the occupancy model of the first robot at the time of conflict and the occupancy model of the second robot at all times, and determine the time when the conflict ends.

[0071] In some embodiments, based on a size conflict index table, the movement lock grid of the first robot at conflict time T is compared with the movement lock grid of the second robot along its entire path, time-step by time, to determine the conflict termination time TE. Simultaneously, the earliest time TS at which the first robot arrives at the node position at conflict time T is recorded.

[0072] Step 54: Based on the earliest time when the first robot arrives at the node where the conflict occurs and the time when the conflict ends, generate all action constraints for the first robot relative to the node where the conflict occurs.

[0073] In some embodiments, a time interval [TS, TE+1] is generated based on the earliest time when the first robot arrives at the node where the conflict occurs and the time when the conflict ends. The time step TE plus 1 is used to prevent following conflicts. All action constraints of the first robot at the node position where the conflict occurs within this time interval [TS, TE+1] are generated and used for subsequent CT child node inheritance.

[0074] Step 55: Based on the size conflict index table, traverse the time steps in reverse with the conflict time as the reference, perform preset operations on the target node where the first robot is located at each time, find the avoidance node, and impose motion constraints on the first robot.

[0075] In some embodiments, see Figure 6 Based on the size conflict index table, the time steps are traversed in reverse order with the conflict time as the reference, and the following operations are performed on the target node where the first robot is located at each time step: Step 551: For each target node at any given time, obtain the set of nodes that are connected to the target node at that time.

[0076] In some embodiments, the node set may include a number of nodes. Specifically, based on the location of the target node on the map, a set of nodes connected to the target node can be found from the map.

[0077] In some embodiments, since the time steps are traversed in reverse based on the conflict time, the time axis corresponding to the first robot is: t=T, T-1, T-2… Then, according to the size conflict index table, the time steps are traversed in reverse based on the conflict time, and the operations of steps 551 to 555 are performed on the target node where the first robot is located at time t.

[0078] In some embodiments, since the node set may contain some original paths, unavoidable points, and node positions on the workbench, these nodes need to be deleted.

[0079] Step 552: Perform robot size conflict detection on the occupancy model of the first robot at the target node and the occupancy model of the second robot at all times, determine the potential conflict termination time, and record the earliest time when the first robot arrives at the target node.

[0080] In some embodiments, the movement lock grid (placement model) of the first robot at the corresponding node at time t is compared with the full path lock grid of the second robot (robot size conflict detection) to determine the potential conflict termination time te. The earliest time ts at which the first robot arrives at the node position at time t is also recorded.

[0081] Step 553: Detect whether there is a conflict between the occupancy model of the first robot from the target node to the node in the node set and the occupancy model of the second robot in the first time interval.

[0082] In some embodiments, the first time interval is the period between the time preceding the conflict time and the time of conflict.

[0083] In some embodiments, if the occupancy model of the first robot from the target node to the node in the node set is found to conflict with the occupancy model of the second robot in the first time interval, step 554 is executed.

[0084] In some embodiments, if the occupancy model of the first robot from the target node to the node in the node set does not conflict with the occupancy model of the second robot in the first time interval, step 555 is executed.

[0085] Step 554: Determine the second time interval based on the earliest time when the first robot arrives at the target node and the potential conflict termination time, and impose motion constraints on the first robot within the second time interval.

[0086] In some embodiments, the time from the earliest moment when the first robot arrives at the target node to the moment when the first potential conflict ends is defined as the second time interval; the moment when the first potential conflict ends is later than the moment when the potential conflict ends; in the second time interval, all actions of the first robot at the target node are constrained.

[0087] Step 555: Determine that there are avoidance nodes in the node set, determine the third time interval based on the earliest time when the first robot arrives at the target node and the potential conflict termination time, and impose action constraints on the first robot in the third time interval.

[0088] In some embodiments, the time from the target moment when the first robot is at the target node to the time when the first potential conflict ends is defined as the third time interval; the target moment is later than the earliest moment when the first robot arrives at the target node, and the time when the first potential conflict ends is later than the time when the potential conflict ends; in the third time interval, the waiting action and moving action of the first robot at the target node are constrained, while the rotation action is retained.

[0089] Step 35: Based on the avoidance nodes and motion constraints, re-plan the paths for the first and second robots until the final planned paths for the first and second robots are obtained.

[0090] In some embodiments, the final planned paths of the first robot and the second robot ensure that their movements do not conflict with each other.

[0091] Compared to the CBS algorithm, which, even with symmetry breaking techniques, still struggles to effectively detect lane size conflicts in densely populated areas, this application adds a lane size conflict detection mechanism to the topology map path planning. By detecting the presence of consecutive conflicts in the immediate vicinity of each other and finding avoidance points for the conflicting robot, constraints are added to guide the robot in planning avoidance actions, thus reducing the recurrence of conflict resolution in space-constrained scenarios. In other words, this application implements lane size conflict detection in topology map planning. For lane size conflicts caused by robot load actions in densely populated areas, traditional CBS algorithms, even with symmetry breaking techniques, often only identify lane size conflicts with matching node positions, failing to effectively identify such conflicts in space-constrained scenarios. This application effectively identifies lane size conflicts, finds avoidance paths for the conflicting robot, reduces the number of conflict resolution attempts, and improves solution efficiency.

[0092] Furthermore, traditional single-machine planning algorithms often use path cost as the sole indicator. This application uses a roadway size conflict index table to guide the pathfinding process of the bounded suboptimal single-machine planning algorithm, thereby reducing the probability of secondary size conflicts.

[0093] In one application scenario, the path planning process is as follows: Step 1: First, determine if the robot's starting point and ending point are on the 2D map. Based on the robot's current position and the coordinates of the task's ending point, find the node closest to it. Then, determine if the distance between this position and the node is less than a preset threshold. If it is less than the preset threshold, it is considered to be on that node; otherwise, a virtual node needs to be created on the map, and the position information, connectivity information, and attribute information between nodes should be synchronized.

[0094] Step 2: Initialize the map and the robot's real-time information, and input the robot's current position and task endpoint information into the MRPP solver.

[0095] Step 3: First, initialize the CT root node. Based on map connectivity and task start point information, plan the optimal path for each robot and break the path down into a time-step sequence. The sequence records the target node, coordinates, angle, and robot actions at each time step, which are used for subsequent collision detection when generating motion lock grids. Perform the first time-step collision detection on the path combination generated by the CT root node: if there is no collision, return the final path directly; otherwise, collision resolution is required.

[0096] Step 4: During conflict resolution, the first step is to determine if there are any unexpanded CT nodes and whether the time has reached a preset threshold. Regarding CT node selection, priority is given to CT nodes with lower conflict counts. If the conflict counts are the same, then CT nodes with lower path costs are further selected. If the final selected CT node has no conflicts, the final path is returned, and the process proceeds to Step 7.

[0097] Step 5: Traverse the conflict information in the CT node, prioritizing lane conflicts, followed by conflicts at earlier time steps. Based on the conflict information, generate constraints of different priorities for the two robots to avoid conflicts, and split into two CT child nodes based on this CT node. The constraint form is as follows: i ,T,action n >, where a i Let represent the i-th robot that needs to execute constraints, T represent the conflict moment, and action. n This represents the set of nth actions constrained at time T. Conflicts are avoided by prohibiting a robot from performing a specific action at time T, such as prohibiting the first robot from performing action at time T. n1 Or prevent the second robot from executing the action at time T. n2 .

[0098] Step 6: Based on the constraint information, replan the path for the robot and perform time-step conflict detection, generating conflict information and storing it in the CT node. Then, jump to step 4 to continue conflict resolution.

[0099] Step 7: After obtaining the final path planned according to MRPP, perform a path distribution safety check. If the check passes, distribute the path to the corresponding robot.

[0100] In some embodiments, collision detection and collision classification are performed in steps 3 and 6. Collision detection is performed in the following order: Step 11: Because collision detection in a topology map differs from point detection in a grid map, it requires not only matching node positions but also detecting the size of the moving lock cells. During collision detection, collisions are detected by traversing the moving lock cells corresponding to each time step of the robot and the moving lock cells of another robot.

[0101] Step 12: When two robots are in conflict, the following two sets of time sequences are used to determine if the node positions are simultaneously identical: (1) The node position of AGV1 at time T-1 and the node position of AGV2 at time T+1. Wherein, AGV1 represents the first robot and AGV2 represents the second robot. ​

[0102] (2) The node position of AGV2 at time T-1 and the node position of AGV1 at time T+1.

[0103] When the conflict node is located at the end point (starting point or ending point) of the path, it can be determined as a roadway conflict if either of the two sets of node position matching is satisfied.

[0104] Step 13: When two robots are in conflict, but it is not a lane conflict, then perform lane size conflict detection. For the conflict time T, verify whether the following two sets of timing sequence movement lock grids simultaneously have secondary conflicts: (1) The movement lock of AGV1 at time T-1 and the movement lock of AGV2 at time T+1.

[0105] (2) The movement and locking of AGV2 at time T-1 and the movement and locking of AGV1 at time T+1.

[0106] When a conflicting node is located at a path endpoint (starting or ending point), if any one of the two sets of moving lock cells satisfying the above two time sequences exists, it can be determined as a roadway size conflict, and the conflict type is updated. This detection process can be understood as a supplement to the time-step moving lock cell detection. Considering that moving lock cell detection takes longer than node position comparison, if the conflict can be determined to be a roadway conflict through node position comparison, then roadway size conflict detection is unnecessary. Figure 7 As shown, AGV1 and AGV2 have a roadway conflict at time T=3. However, the roadway conflict detection method based on node position matching cannot determine this type of conflict. Since there is a movement lock grid conflict before and after the conflict time T, it is determined to be a conflict with opposite roadway dimensions.

[0107] In some embodiments, in step 5, the conflicts to be resolved are selected. When the conflict to be resolved is a roadway size conflict, corresponding constraints need to be generated for the roadway size conflict. The specific steps are as follows: Step 1: Perform bidirectional recursive detection based on the conflict time T: Perform movement lock detection on AGV1 in descending time step order (t=T, T-1, T-2…) and AGV2 in ascending time step order (t=T, T+1, T+2…), and then perform detection on AGV1 in ascending order (t=T, T+1, T+2…) and AGV2 in descending order (t=T, T-1, T-2…), recursively until no new conflicts are generated. Finally, generate a roadway size conflict index table. i , <I,<N i K i T i >>>and j , <I,<N j K j T​​j >>>, where 'a' is the AGV number. <N i K i T i >For AGV i The relevant conflict information is as follows: I is the lane conflict index, N is the conflict node location, K is the index of the moving lock grid in the corresponding AGV path, and T is the time when the conflict occurred.

[0108] Step 2: When AGV1 is required to avoid AGV2, a waiting action is generated for AGV2 at the node position at time T-1. This action, along with all possible actions of AGV1 at time T-1, is then subjected to a lock-by-lock conflict detection with AGV2. Detected conflicts are converted into constraints and added to the constraint set of AGV1 for subsequent CT child node inheritance. By pre-detecting the action of the node before the conflict occurs, a diagonal avoidance path is generated to eliminate unilateral avoidance size conflicts.

[0109] Step 3: Since conflict resolution needs to consider following conflicts, it's necessary to avoid continuous conflict resolution at the moment of conflict. Based on the lane size conflict index table, the movement lock of AGV1 at the conflict time T is compared with the movement lock of AGV2 along the entire path step by step to determine the conflict termination time TE. Simultaneously, the earliest time TS when AGV1 arrives at the node position at the conflict time T is recorded. The generated time interval is [TS, TE+1], where TE plus 1 time step is to prevent following conflicts. All action constraints of AGV1 at the node position at the conflict time T within this time interval [TS, TE+1] are generated and used for subsequent CT child node inheritance.

[0110] Step 4: Based on the lane size conflict index table, traverse the time axis in reverse order starting from the conflict time T (t=T, T-1, T-2…), and execute the following steps for the node of AGV1 at time t: (1) Obtain the set of nodes with connectivity {neighbor(n)|n∈1,2,3…}.

[0111] (2) Filter out the nodes in neighbor(n) that are located on the original paths of the two robots, unavoidable points, and workbench locations.

[0112] (3) Check whether there is a conflict between the movement lock of AGV1 from the node to neighbor(n) and the movement lock of AGV2 in the time interval [T-1,T]. If there is a conflict, add a movement lock constraint to AGV1; if there is no conflict, it is considered that there is a valid avoidance action.

[0113] (4) Compare the movement lock grid of AGV1 at the corresponding node at time t with the full path lock grid of AGV2 to determine the potential conflict termination time te. And record the earliest time ts when AGV1 arrives at the node position at time t.

[0114] (5) Through the conflict detection in (3), if there is no effective avoidance action, add all action constraints for AGV1 at the node position at time t in the time interval [ts, te+1]; otherwise, add waiting and moving action constraints for AGV1 at the node position at time t in the time interval [ts+1, te+1], retain the rotation action to adjust the direction, and use it for subsequent CT child node inheritance. Figure 8 As shown, AGV1 finds an avoidance path at avoidance node K, and its arrival time at node A is t=1. Based on the lane size conflict index table, AGV2 and AGV1 collide at node A at time t=4. Therefore, the earliest time AGV1 arrives at node A should be t=1, and its departure time should be t=2. This constraint duration should continue until AGV2 resolves the conflict with AGV1's movement lock at node A. If the neighbor(n) set has been traversed and no safe avoidance action is found, then Step 4 is repeated.

[0115] In some embodiments, in step 6, the low-level single-machine planning algorithm prioritizes the shortest path based on constraints. However, excessively short constraint durations can lead to ineffective selection of avoidance paths. Therefore, a bounded suboptimal search algorithm is introduced. This algorithm guides the pathfinding process of the single-machine planning algorithm through a lane size conflict index table. During node expansion, a set of candidate nodes with path lengths within the suboptimal range is selected. Priority is given to selecting the node with the fewest overlapping nodes with other robot paths at the current moment, followed by the node with the fewest overlapping nodes with size conflict nodes. [admin7] See Figure 9 , Figure 9 This is a schematic diagram of an embodiment of the robot path planning device provided in this application. The robot path planning device 30 includes a processor 301, a memory 302, and a communication module 303. The memory 302 is used to store computer programs, and the processor 301 is used to execute the computer programs to implement the following methods: Path planning is performed on the first robot and the second robot to obtain a first planned path for the first robot and a second planned path for the second robot. Based on the first planned path and the second planned path, mutual motion conflict detection is performed on the first robot and the second robot. Mutual motion conflict detection includes robot size conflict detection during mutual motion. In response to the mutual motion conflict between the first robot and the second robot, an avoidance node is found for the first robot or the second robot, and motion constraints are applied to the first robot or the second robot. Based on the avoidance node and motion constraints, path planning is performed again for the first robot and the second robot until the final planned path for the first robot and the second robot is obtained.

[0116] In some embodiments, the processor 301 is further configured to execute a computer program to implement the following method: in response to a conflict of movement between the first robot and the second robot, obtaining the conflict time corresponding to the conflict of movement; finding an avoidance node for the first robot or the second robot according to the conflict time, and imposing motion constraints on the first robot or the second robot.

[0117] In some embodiments, the processor 301 is further configured to execute a computer program to implement the following method: backtracking the first planned path and the second planned path step by step based on the moment of conflict; during the step-by-step backtracking process, finding avoidance nodes based on node positions and path information, and imposing motion constraints on the first robot or the second robot.

[0118] In some embodiments, the processor 301 is further configured to execute a computer program to implement the following method: performing robot size conflict detection on the first robot in descending order of time steps and on the second robot in ascending order of time steps, based on the conflict time; and performing robot size conflict detection on the first robot in ascending order of time steps and on the second robot in descending order of time steps, based on the conflict time, to obtain a size conflict index table corresponding to the first robot and the second robot; in response to the first robot needing to avoid the second robot, generating a waiting action for the second robot at the node corresponding to the previous time step at the conflict time, and comparing all actions of the first robot at the previous time step with... The system waits for robot size conflict detection to be performed, generating motion constraints for the first robot. Based on the size conflict index table, it performs robot size conflict detection between the occupant model of the first robot at the time of conflict and the occupant model of the second robot at all times, determining the conflict termination time. Based on the earliest time when the first robot arrives at the node where the conflict occurs and the conflict termination time, it generates all motion constraints for the first robot relative to the node where the conflict occurs. Based on the size conflict index table, it traverses the time steps in reverse with the conflict time as the reference, performs preset operations on the target node where the first robot is located at each time, finds avoidance nodes, and applies motion constraints to the first robot.

[0119] In some embodiments, the processor 301 is further configured to execute a computer program to implement the following method: according to a size conflict index table, traversing time steps in reverse with the conflict time as a reference, performing the following operations on the target node where the first robot is located at each time: for the target node at each time, obtaining a set of nodes that are connected to the target node at each time; and performing robot size conflict detection on the occupancy model of the first robot at the target node and the occupancy model of the second robot at all times, determining the potential conflict termination time, and recording the earliest time when the first robot arrives at the target node; detecting whether there is a conflict between the occupancy model of the first robot from the target node to the node in the node set and the occupancy model of the second robot in a first time interval; the first time interval is between the time before the conflict time and the conflict time; if yes, determining a second time interval based on the earliest time when the first robot arrives at the target node and the potential conflict termination time, and imposing action constraints on the first robot in the second time interval; if no, determining that there is an avoidance node in the node set, determining a third time interval based on the earliest time when the first robot arrives at the target node and the potential conflict termination time, and imposing action constraints on the first robot in the third time interval.

[0120] In some embodiments, the processor 301 is further configured to execute a computer program to implement the following method: determining a second time interval from the earliest time when the first robot arrives at the target node to the time when the first potential conflict ends; the time when the first potential conflict ends is later than the time when the potential conflict ends; and constraining all actions of the first robot at the target node during the second time interval.

[0121] In some embodiments, the processor 301 is further configured to execute a computer program to implement the following method: determining a third time interval from the target time when the first robot is at the target node to the first potential conflict termination time; the target time is later than the earliest time when the first robot arrives at the target node, and the first potential conflict termination time is later than the potential conflict termination time; in the third time interval, constraining the waiting and moving actions of the first robot at the target node, while retaining the rotation action.

[0122] In some embodiments, the processor 301 is further configured to execute a computer program to implement the method of any embodiment of this application.

[0123] See Figure 10 , Figure 10 This is a schematic diagram of an embodiment of the computer-readable storage medium provided in this application. The computer-readable storage medium 200 stores a computer program 201, which, when executed by a processor, implements the following method: Path planning is performed on the first robot and the second robot to obtain a first planned path for the first robot and a second planned path for the second robot. Based on the first planned path and the second planned path, mutual motion conflict detection is performed on the first robot and the second robot. Mutual motion conflict detection includes robot size conflict detection during mutual motion. In response to the mutual motion conflict between the first robot and the second robot, an avoidance node is found for the first robot or the second robot, and motion constraints are applied to the first robot or the second robot. Based on the avoidance node and motion constraints, path planning is performed again for the first robot and the second robot until the final planned path for the first robot and the second robot is obtained.

[0124] In some embodiments, when the computer program 201 is executed by the processor, it is further configured to implement the following method: in response to a conflict between the first robot and the second robot, obtain the conflict time corresponding to the conflict; find an avoidance node for the first robot or the second robot according to the conflict time, and impose motion constraints on the first robot or the second robot.

[0125] In some embodiments, when the computer program 201 is executed by the processor, it is also used to implement the following method: backtracking the first planned path and the second planned path step by step based on the moment of conflict; during the backtracking process step by step, finding avoidance nodes based on node positions and path information, and imposing action constraints on the first robot or the second robot.

[0126] In some embodiments, when executed by a processor, the computer program 201 is further configured to implement the following method: performing robot size conflict detection on the first robot in descending order of time steps and on the second robot in ascending order of time steps, based on the conflict time; and performing robot size conflict detection on the first robot in ascending order of time steps and on the second robot in descending order of time steps, based on the conflict time, to obtain a size conflict index table corresponding to the first robot and the second robot; in response to the first robot needing to avoid the second robot, generating a waiting action for the second robot at the node corresponding to the previous time step at the conflict time, and retrieving all actions of the first robot at the previous time step. Robot size conflict detection is performed on the waiting action to generate motion constraints for the first robot; based on the size conflict index table, robot size conflict detection is performed between the occupant model of the first robot at the conflict moment and the occupant model of the second robot at all moments to determine the conflict termination moment; based on the earliest moment when the first robot arrives at the node where the conflict moment is located and the conflict termination moment, all motion constraints for the first robot relative to the node where the conflict moment is located are generated; based on the size conflict index table, the time steps are traversed in reverse with the conflict moment as the reference, and preset operations are performed on the target node where the first robot is located at each moment to find avoidance nodes and to impose motion constraints on the first robot.

[0127] In some embodiments, when the computer program 201 is executed by the processor, it is further configured to implement the following method: according to the size conflict index table, traversing the time steps in reverse with the conflict time as the reference, performing the following operations on the target node where the first robot is located at each time: for the target node where the first robot is located at each time, obtaining a set of nodes that are connected to the target node where the first robot is located at each time; and performing robot size conflict detection on the occupancy model of the first robot at the target node and the occupancy model of the second robot at all times, determining the potential conflict termination time, and recording the earliest time when the first robot arrives at the target node; detecting whether there is a conflict between the occupancy model of the first robot from the target node to the node in the node set and the occupancy model of the second robot in the first time interval; the first time interval is from the time before the conflict time to the conflict time; if yes, determining the second time interval based on the earliest time when the first robot arrives at the target node and the potential conflict termination time, and imposing action constraints on the first robot in the second time interval; if no, determining that there is an avoidance node in the node set, determining the third time interval based on the earliest time when the first robot arrives at the target node and the potential conflict termination time, and imposing action constraints on the first robot in the third time interval.

[0128] In some embodiments, when the computer program 201 is executed by the processor, it is further configured to implement the following method: determining a second time interval from the earliest time when the first robot arrives at the target node to the time when the first potential conflict ends; the time when the first potential conflict ends is later than the time when the potential conflict ends; and constraining all actions of the first robot at the target node during the second time interval.

[0129] In some embodiments, when the computer program 201 is executed by the processor, it is further configured to implement the following method: determining the target time from the time when the first robot is at the target node to the time when the first potential conflict ends as a third time interval; the target time is later than the earliest time when the first robot arrives at the target node, and the time when the first potential conflict ends is later than the time when the potential conflict ends; in the third time interval, constraining the waiting action and moving action of the first robot at the target node, while retaining the rotation action.

[0130] In some embodiments, when the computer program 201 is executed by a processor, it is also used to implement the method of any embodiment of this application.

[0131] In summary, the robot path planning method, apparatus, system, and readable storage medium provided in this application, after performing path planning on a first robot and a second robot to obtain a first planned path for the first robot and a second planned path for the second robot, performs mutual motion conflict detection on the first robot and the second robot based on the first planned path and the second planned path. For example, mutual motion conflict detection includes robot size conflict detection during mutual motion. In this way, when there is mutual motion conflict between the first robot and the second robot, an avoidance node is found for the first robot or the second robot, and motion constraints are applied to the first robot or the second robot. Based on the avoidance node and motion constraints, the path planning of the first robot and the second robot is re-performed until the final planned path of the first robot and the second robot is obtained. This can effectively identify size-related mutual conflicts in narrow spaces such as alleyways, find avoidance paths for the conflicting first robot and the second robot, reduce the number of times such conflicts are resolved, and improve solution efficiency.

[0132] Furthermore, compared to the CBS algorithm which, even with symmetry breaking techniques, cannot effectively detect lane size conflicts between robots in densely populated areas, this application adds a lane size conflict detection mechanism to the topology map path planning. By detecting whether there are consecutive conflicts in the immediate vicinity of the conflict, it finds an avoidance point for the conflicting robot and guides the robot to plan avoidance actions by adding constraints, thereby reducing the continuous occurrence of conflict resolution in space-constrained scenarios. In other words, this application implements lane size conflict detection functionality in topology map planning. For lane size conflicts caused by robot load actions in densely populated areas in real-world scenarios, traditional CBS algorithms, even with symmetry breaking techniques, often only identify lane size conflicts with matching node positions, making it difficult to effectively identify such conflicts in space-constrained scenarios. This application can effectively identify lane size conflicts, find avoidance paths for the conflicting robot, reduce the number of conflict resolution attempts, and improve solution efficiency.

[0133] Furthermore, traditional single-machine planning algorithms often use path cost as the sole indicator. This application uses a roadway size conflict index table to guide the pathfinding process of the bounded suboptimal single-machine planning algorithm, thereby reducing the probability of secondary size conflicts.

[0134] In the several embodiments provided in this application, it should be understood that the disclosed methods and devices can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of circuits or units is merely a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0135] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0136] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0137] 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 changes made based on the description and drawings of this application, 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 planning method for a robot, characterized in that, The robot includes at least a first robot and a second robot, and the path planning method includes: Path planning is performed on the first robot and the second robot to obtain a first planned path for the first robot and a second planned path for the second robot. Based on the first planned path and the second planned path, the first robot and the second robot are subjected to mutual motion conflict detection; the mutual motion conflict detection includes robot size conflict detection during mutual motion. In response to a conflict of movement between the first robot and the second robot, the moment of the conflict is obtained. Based on the conflict moment, backtrack the first planned path and the second planned path step by step; During the time-step backtracking process, avoidance nodes are found based on node positions and path information, and action constraints are imposed on the first robot or the second robot. Based on the avoidance node and the action constraint, re-plan the path for the first robot and the second robot until the final planned path for the first robot and the second robot is obtained; The step of finding avoidance nodes based on node positions and path information and imposing action constraints on the first robot during the time-step backtracking process includes: Based on the conflict time, robot size conflict detection is performed on the first robot in descending order of time steps and on the second robot in ascending order of time steps; and based on the conflict time, robot size conflict detection is performed on the first robot in ascending order of time steps and on the second robot in descending order of time steps, to obtain a size conflict index table corresponding to the first robot and the second robot. In response to the first robot needing to avoid the second robot, a waiting action is generated for the second robot at the node corresponding to the previous moment at the time of the conflict, and the robot size conflict detection is performed on all the actions of the first robot at the previous moment and the waiting action to generate action constraints for the first robot. According to the size conflict index table, the occupancy model of the first robot at the time of the conflict and the occupancy model of the second robot at all times are used to perform robot size conflict detection to determine the conflict termination time. Based on the earliest time when the first robot arrives at the node where the conflict occurs and the time when the conflict ends, generate all action constraints of the first robot relative to the node where the conflict occurs. Based on the size conflict index table, the time steps are traversed in reverse with the conflict time as the reference. A preset operation is performed on the target node where the first robot is located at each time to find the avoidance node and to constrain the action of the first robot.

2. The path planning method according to claim 1, characterized in that, The step of reversely traversing time steps based on the size conflict index table and the conflict time, performing preset operations on the target node where the first robot is located at each time step, finding the avoidance node, and imposing motion constraints on the first robot includes: Based on the size conflict index table, and using the conflict time as a reference, the time steps are traversed in reverse order, and the following operations are performed on the target node where the first robot is located at each time step: For each target node at any given time, obtain the set of nodes that are connected to the target node at that time. And perform robot size conflict detection on the occupancy model of the first robot at the target node and the occupancy model of the second robot at all times, determine the potential conflict termination time, and record the earliest time when the first robot arrives at the target node; Detect whether there is a conflict between the placeholder model of the first robot from the target node to the node in the node set and the placeholder model of the second robot in a first time interval; the first time interval is from the moment before the conflict moment to the conflict moment. If so, a second time interval is determined based on the earliest time when the first robot arrives at the target node and the potential conflict termination time, and the first robot is subject to action constraints during the second time interval; If not, determine that the avoidance node exists in the node set, determine a third time interval based on the earliest time when the first robot arrives at the target node and the potential conflict termination time, and impose action constraints on the first robot in the third time interval.

3. The path planning method according to claim 2, characterized in that, The step of determining a second time interval based on the earliest time the first robot arrives at the target node and the potential conflict termination time, and imposing action constraints on the first robot within the second time interval, includes: The second time interval is defined as the time from the earliest time the first robot arrives at the target node to the time when the first potential conflict ends; the time when the first potential conflict ends is later than the time when the potential conflict ends. During the second time interval, constraints are applied to all actions of the first robot at the target node.

4. The path planning method according to claim 2, characterized in that, The step of determining a third time interval based on the earliest time the first robot arrives at the target node and the potential conflict termination time, and imposing action constraints on the first robot within the third time interval, includes: The third time interval is defined as the time from the target moment when the first robot is at the target node to the time when the first potential conflict ends; the target moment is later than the earliest moment when the first robot arrives at the target node, and the time when the first potential conflict ends is later than the time when the potential conflict ends. During the third time interval, the waiting and moving actions of the first robot at the target node are constrained, while the rotation action is retained.

5. A path planning device for a robot, characterized in that, The path planning device of the robot includes a processor, a memory, and a communication module. The memory is used to store a computer program, and the processor is used to execute the computer program to implement the method as described in any one of claims 1-4.

6. A robot system, characterized in that, The robot system includes: Multiple robots, including at least a first robot and a second robot; A path planning device for a robot, as described in claim 5.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, is used to implement the method as described in any one of claims 1-4.

Images