Multi-robot layered space-time optimization path planning method based on conflict resolution

Abstract

The invention discloses a multi-robot layered space-time optimization path planning method based on conflict resolution, and relates to the technical field of multi-robot path planning, and the method comprises the steps: constructing a semantic topology and a grid hierarchical graph of a multi-robot operation environment, and obtaining a robot preset path; robot path conflicts are detected and classified; calculating corresponding dynamic priorities of the robots with the path conflicts; evaluating the total cost of each conflict resolution strategy of the robot path conflict; obtaining an optimal conflict elimination strategy of each robot according to a principle of minimum total cost; according to the corresponding dynamic priorities of the robots, to-be-processed robots are screened out; and performing path re-planning on the to-be-processed robot based on the optimal conflict elimination strategy of the to-be-processed robot and a preset path. According to the method, the robots with low priorities can be re-planned with the lowest cost, and finally robot conflicts are eliminated.

Description

Technical Field

[0001] The present invention relates to the technical field of multi-robot path planning, and in particular to a multi-robot hierarchical spatiotemporal optimization path planning method based on conflict resolution. Background Art

[0002] With the rapid development of the national economy, the demand for family cars in my country continues to grow. More and more families own family cars, and the number of private cars has increased dramatically. According to relevant agencies, my country's car ownership will rank first in the world, but correspondingly, there will be an increasing gap in the demand for parking spaces in my country. The difficulty of parking has brought a series of problems to the people: the shortage of parking spaces in the city has forced some car owners to park on both sides of the road, which not only affects the beauty of the city but also worsens the traffic conditions in the city.

[0003] With the rapid development of domestic robot technology, the intelligent parking system based on autonomous parking robots has gradually become an effective solution to the parking problem. Autonomous parking robots are automated parking devices that move vehicle-carrying equipment to the target location. The entire parking process is unmanned. When parking, there is no need to reserve space for passengers to get on and off the car, which increases the utilization rate of land in disguise. According to statistics, parking can increase the parking capacity by more than 30%. At the same time, unmanned parking means that problems such as manual parking and finding parking spaces are reduced, which also greatly improves parking efficiency.

[0004] However, due to the limited working capacity of a single robot, it is difficult to complete complex tasks. Intelligent parking systems often require multiple robots to work together to complete vehicle storage and parking tasks. Multi-robot path planning is much more complicated than single-robot path planning. When there are multiple mobile robots in a scene, even if each robot has an obstacle avoidance strategy, it is easy to cause road congestion and blockage. The multi-robot path planning system must not only generate an optimized path from the starting point to the target point for a single robot, but also avoid conflicts and collisions between the robot and surrounding obstacles and other robots, avoid deadlock problems, and enable multiple robots to successfully complete the specified task through collaboration. Since parking environments often have random tasks and dynamic adjustments, such as the movement of other vehicles and the occupation of parking spaces, etc. The path planning algorithm needs to be able to accurately and in real time perceive and update map data, and incorporate it into path planning to make dynamic adjustments to path planning. At intersections, multiple robots may arrive at or approach the intersection at the same time. It is necessary to consider the yield rules between robots, determine the order of robots' passage, and allocate appropriate time intervals to avoid intersection congestion and accidents. Finally, the algorithm needs to generate high-quality paths within a limited time to meet real-time requirements.

[0005] Currently, no effective solution has been proposed for the problems in the related technologies. Summary of the invention

[0006] In view of the problems in the related art, the present invention proposes a multi-robot hierarchical spatiotemporal optimization path planning method based on conflict resolution to overcome the above-mentioned technical problems existing in the existing related technology.

[0007] To this end, the specific technical solution adopted by the present invention is as follows:

[0008] A multi-robot hierarchical spatiotemporal optimization path planning method based on conflict resolution, the method comprising:

[0009] S1. Constructing a semantic topology and a grid hierarchical graph of a multi-robot operation environment, and using a graph search algorithm in the semantic topology and the grid hierarchical graph to generate a preset robot path from a start node to a target node;

[0010] S2. Based on the pre-configured classification rules, the robot path conflict type is preliminarily detected, and the pre-trained deep learning model is used to correct the preliminary detection of the robot path conflict type to obtain the detection and classification results of the robot path conflict;

[0011] S3, calculating the corresponding dynamic priority of the robot with path conflict according to the entanglement degree of the robot path and the path execution time;

[0012] S4. Based on the detection and classification results of the robot path conflict, the pre-established conflict type resolution strategy template library and the cost function, the total cost of each conflict resolution strategy of the robot path conflict is evaluated; according to the principle of minimum total cost, the optimal conflict elimination strategy for each robot is obtained;

[0013] S5. Filter out the robots to be processed according to their corresponding dynamic priorities; and replan the paths of the robots to be processed based on the optimal conflict elimination strategies and preset paths of the robots to be processed.

[0014] Furthermore, a semantic topology and a grid hierarchical graph of a multi-robot operation environment are constructed, and a robot preset path from a start node to a target node is generated using a graph search algorithm in the semantic topology and the grid hierarchical graph, including:

[0015] Import the working environment grid map and element information, and perform spatiotemporal gridding on the map. The robot width information and the safety distance between robots are used as the spatial resolution of the working environment grid map, and the time interval of the grid occupied is determined according to the robot driving speed.

[0016] Label and cluster the semantic information of objects in the robot's working environment, and form a semantic road network diagram of the working environment based on the semantic information of objects;

[0017] The semantic road network map of the working environment is merged with the grid map of the working environment to obtain the semantic topology and grid hierarchical map. In the semantic topology and grid hierarchical map, the shortest path from the start node to the target node generated based on the graph search algorithm is used as the preset path of the robot.

[0018] Furthermore, based on the pre-configured classification rules, the robot path conflict type is initially detected, and the pre-trained deep learning model is used to correct the initial detection of the robot path conflict type, and the robot path conflict detection and classification results are obtained, including:

[0019] Using pre-configured classification rules, obtain preliminary detection and classification results of robot path conflict types;

[0020] The environmental perception module is used to obtain the environmental perception sensor data of the multi-robot working environment, and combined with the robot's state information to obtain input data, and obtain different types of robot conflict labels as output data;

[0021] Train the pre-configured multi-layer MLP model using input data and output data, and optimize the parameters of the multi-layer MLP model;

[0022] Using the detection and classification results output by the multi-layer MLP model, the preliminary detection and classification results are corrected by weighted adjustment to obtain the detection and classification results of the robot path conflict.

[0023] Furthermore, using the pre-configured classification rules, the preliminary detection and classification results of the robot path conflict types are obtained, including:

[0024] When the paths of the robots intersect, the type of robot path conflict is determined to be a fixed-point conflict;

[0025] When the robot group crosses the edge and exchanges nodes at the same time, the type of robot path conflict is judged to be fixed-point conflict;

[0026] When the node where the robot is located is occupied at the next moment, the type of robot path conflict is determined to be a following conflict;

[0027] When the robot group forms a ring, the type of robot path conflict is judged to be a ring conflict.

[0028] Furthermore, according to the entanglement degree of the robot path and the path execution time, the dynamic priority of the robot with path conflict is calculated including:

[0029] Obtain the total number of conflicts between each robot path and other paths, and assign priorities to robots based on the principle that the greater the total number of conflicts, the lower the priority of the robot;

[0030] When the total number of conflicts in the robot group is the same, the priority of the robots is assigned according to the principle that the shorter the path execution time, the lower the priority of the robot;

[0031] When the total number of conflicts and execution time of the robot group are the same, the avoidance party and the priority party are randomly selected.

[0032] Furthermore, the pre-established conflict type resolution strategy template library includes:

[0033] According to the different conflict types of robots, formulate corresponding conflict resolution strategies; based on the conflict resolution strategies, establish a conflict type resolution strategy template library.

[0034] Furthermore, based on the detection and classification results of the robot path conflict, the pre-established conflict type resolution strategy template library and the cost function, the total cost of each conflict resolution strategy of the robot path conflict is evaluated, including:

[0035] By defining different robot path conflict types and conflict type resolution strategies, a cost function is established; the weights in the cost function are adjusted according to the pre-set logic;

[0036] Based on the detection and classification results of the robot path conflict and the pre-established conflict type resolution strategy template library, several conflict type resolution strategies are obtained; and the total cost corresponding to each conflict type resolution strategy is calculated using the cost function.

[0037] Furthermore, the weights in the cost function are adjusted according to a pre-set logic including:

[0038] When the obstacle density is higher than the preset obstacle density threshold, the safety weight is increased;

[0039] When the obstacle density is lower than the preset obstacle density threshold, the path efficiency weight is increased;

[0040] When the robot density on the robot path is higher than the preset robot density threshold, the avoidance and coordination weights are increased;

[0041] When the robot density on the robot path is lower than the preset robot density threshold, the speed weight is increased;

[0042] When an emergency is detected, the weighting of safety and response speed is increased.

[0043] Furthermore, according to the corresponding dynamic priorities of the robots, the robots to be processed are screened out, including:

[0044] The dynamic priorities of the robots are compared, and the robot with the lowest priority is selected as the robot to be processed.

[0045] Furthermore, based on the optimal conflict elimination strategy and preset path of the robot to be processed, path replanning of the robot to be processed includes:

[0046] Obtain a preset path of the robot to be processed; and adjust the preset path of the robot to be processed to a replanned path by executing an optimal conflict elimination strategy of the robot to be processed.

[0047] The beneficial effects of the present invention are:

[0048] (1) The present invention constructs a semantic road network graph-grid subgraph to store the semantic constraints, geometric constraints and strategy constraints in the working environment in the form of a database, thereby simplifying the feasible area of ​​the robot's path planning; the grid map is divided into regions based on the semantic constraints of objects in the graph model, so that subsequent path conflict resolution is redistributed in the spatiotemporal grid subgraph, which simplifies the optimization space of the algorithm and improves the real-time performance of the algorithm.

[0049] (2) The present invention adopts a hierarchical space-time optimization path planning algorithm, and balances the real-time nature of the calculation and the existence of the solution through the strategy of space-time decoupled spatial path conflict screening-space-time joint conflict resolution. The space-time decoupled spatial path conflict screening algorithm first detects the conflict in this situation and classifies the conflict, and then uses the space-time joint conflict resolution algorithm to plan a conflict-free path. Compared with multiple types of spatial planning algorithms such as reserved areas and A*, when the number of robots is too large, invalid situations where the robot path does not exist are prone to occur; the space-time joint planning algorithm ensures the existence of multi-robot solutions, that is, in extreme cases, the robots can be executed sequentially by single machines with time redistribution. The method of the present invention reduces the number of node expansions at the high level of the algorithm and the number of low-level path re-planning times, is not affected by the operating environment and the number of robots, and can ensure that multiple robots complete the task.

[0050] (3) The present invention classifies path conflicts more clearly, proposes key opposite vertex conflicts, key intersection vertex conflicts and key edge conflicts, and designs solutions for conflict types. The method of the present invention adds constraints according to the conflict type at a high level, and updates the set of shortest paths that satisfy the constraints at a low level by replanning the path or waiting. The method of the present invention can resolve the current conflict without causing new avoidable conflicts, and can resolve multiple conflicts with one path update. The method of the present invention effectively reduces the overall computational complexity of the algorithm and can quickly find a suboptimal solution for the multi-robot path planning problem. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0052] Figure 1 is a flow chart of a multi-robot hierarchical spatiotemporal optimization path planning method based on conflict resolution according to an embodiment of the present invention;

[0053] Figure 2 A diagram of a multi-robot path planning system with hierarchical spatiotemporal optimization based on conflict resolution;

[0054] Figure 3 This is a schematic diagram of a semantic road network-grid hierarchical map;

[0055] Figure 4 It is a schematic diagram of the mapping between conflict types and resolution strategies;

[0056] Figure 5 Flowchart of the multi-robot path planning algorithm. DETAILED DESCRIPTION

[0057] To further illustrate each embodiment, the present invention provides drawings, which are part of the disclosure of the present invention and are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these contents, ordinary technicians in the field should be able to understand other possible implementations and advantages of the present invention. The components in the figures are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0058] According to an embodiment of the present invention, a multi-robot hierarchical spatiotemporal optimization path planning method based on conflict resolution is provided. The multi-robot autonomous parking path planning problem is often complex, non-deterministic, and may involve large-scale data and calculations. The efficiency and real-time performance of the algorithm is a challenge, and it is necessary to improve the performance of the path planning algorithm through efficient search and optimization techniques, so that the autonomous parking system can be widely used in various parking lots, including commercial centers, public buildings, hospitals, schools, etc., to bring greater convenience and comfort experience to people's travel.

[0059] The present invention is further described with reference to the accompanying drawings and specific embodiments. Figure 1 As shown, a multi-robot hierarchical spatiotemporal optimization path planning method based on conflict resolution according to an embodiment of the present invention includes:

[0060] S1. Constructing a semantic topology and a grid hierarchical graph of a multi-robot operation environment, and using a graph search algorithm in the semantic topology and the grid hierarchical graph to generate a preset robot path from a start node to a target node;

[0061] S2. Based on the pre-configured classification rules, the robot path conflict type is preliminarily detected, and the pre-trained deep learning model is used to correct the preliminary detection of the robot path conflict type to obtain the detection and classification results of the robot path conflict;

[0062] S3, calculating the corresponding dynamic priority of the robot with path conflict according to the entanglement degree of the robot path and the path execution time;

[0063] S4. Based on the detection and classification results of the robot path conflict, the pre-established conflict type resolution strategy template library and the cost function, the total cost of each conflict resolution strategy of the robot path conflict is evaluated; according to the principle of minimum total cost, the optimal conflict elimination strategy for each robot is obtained;

[0064] S5. Filter out the robots to be processed according to their corresponding dynamic priorities; and replan the paths of the robots to be processed based on the optimal conflict elimination strategies and preset paths of the robots to be processed.

[0065] In order to facilitate understanding of the above technical solutions of the present invention, the above technical solutions of the present invention are further described below from the perspective of architecture and principle, as follows:

[0066] The present invention can provide a feasible conflict-free path for a multi-robot parking system by considering optimality. First, a semantic topology-grid hierarchical map is constructed. The semantic topology map stores the robot's driving key points and semantic constraints, as well as the offline empirical optimal path of the corresponding grid subgraph. When a multi-robot task is released, a hybrid A* is used at the bottom layer to search the grid for trajectories and obtain the initial trajectory, thereby improving the rationality of planning while ensuring the efficiency of trajectory search. The overlapping areas of each robot in space are determined by the robot's pre-planned trajectory, and the extraction of the robot's minimum interference semantic subgraph is realized. Conflict detection and classification are performed in the lower grid map corresponding to the semantic subgraph. The dynamic priority of the conflicting multi-robots is calculated. Based on the conflict detection type and robot priority, the cost of conflict resolution strategies such as waiting in place, driving perpendicular to the original direction, and driving in the opposite direction is evaluated, and the robot with low priority is replanned with the lowest cost in the grid subgraph, and finally the robot conflict is eliminated. The cost evaluation of the conflict resolution strategy is to define different conflict types and their resolution strategies to establish a cost function:

[0067] C i =w1·T i +w2·E i+w3·D i

[0068] Among them, T i is the time cost, E i is the energy cost, D i is the path deviation degree, w1, w2 and w3 are weight coefficients of each item. In addition, the above cost function is not the only cost function.

[0069] After identifying the environment, the environmental feature recognition module adjusts the weights based on the environment complexity, path congestion, real-time changes, and task priority logic: when the obstacle density is high, the safety weight is increased to prioritize obstacle avoidance; when obstacles are sparse, the path efficiency weight is increased to shorten the travel time; when the robots on the path are dense, the avoidance and coordination weights are increased to reduce conflicts; when the path is idle, the speed weight is increased. When an emergency is detected, the safety and response speed weights are increased; if the task priority is high, the task completion is prioritized. Specifically:

[0070] When the obstacle density is higher than the preset obstacle density threshold, the safety weight is increased [for example, when the obstacle density > 0.6 (the number of obstacles per square meter exceeds 0.6), the safety weight is increased to 1.5 times].

[0071] When the obstacle density is lower than the preset obstacle density threshold, the path efficiency weight is increased (for example, when the obstacle density is ≤0.3, the path efficiency weight is increased to 1.2 times).

[0072] When the robot density on the robot path is higher than the preset robot density threshold, the avoidance and coordination weights are increased (for example, when the robot density on the path > the total number of robots*0.7, the avoidance and coordination weights are increased to 2 times).

[0073] When the robot density on the robot path is lower than the preset robot density threshold, the speed weight is increased (for example, when the robot density on the path is less than the total number of robots*0.3, the speed weight is increased to 1.5 times).

[0074] When an unexpected situation is detected, the safety and response speed weights are increased [for example, when an obstacle suddenly appears or the robot fails, the safety weight is immediately increased to the highest level (weight factor 2), and the response speed weight is increased to 1.5 times to ensure that the system can respond quickly and adjust the path].

[0075] At the same time, when the environment is stable, the path smoothness weight is increased to reduce steering and braking, achieving efficient and stable operation. To adapt to different environmental changes, the strategy with the lowest total cost is selected, and real-time feedback is provided to continuously optimize system efficiency and robustness. The robots with low priority are replanned at the lowest cost in the grid subgraph. The overall hierarchical spatiotemporal optimization path planning system based on conflict resolution, such as Figure 2 As shown, the semantic road network-grid hierarchical map, such as Figure 3 As shown, the conflict type and resolution strategy are mapped, such as Figure 4 As shown in Figure 2, the multi-robot path planning algorithm process is as follows: Figure 5 As shown in the figure, a multi-robot hierarchical spatiotemporal optimization path planning method based on conflict resolution is proposed. Figure 5 It includes: calling the semantic road network diagram - generating the shortest path of a single robot Hybrid A* - detecting and classifying path conflicts. When there is a conflict, the conflict robot priority is calculated - calling the conflict type strategy template - dynamically replanning the motion trajectory - calculating the strategy cost, selecting the optimal resolution strategy - dynamically replanning the motion trajectory.

[0076] The specific steps of the present invention are as follows:

[0077] (1) Import the grid map of the working environment and information on elements such as objects, entrances and exits, parking locations, and channel characteristics, and perform spatiotemporal gridding on the map. Assume that robot i is defined, i = 1, 2, 3, ..., m, where m is the total number of robots with the same length and width (l, w). The smaller value width information and the safe distance between robots are used as the spatial resolution of the grid map. The time interval Δt of the grid occupied is determined according to the robot's driving speed. The state of the center point of the robot's rear axle at k time steps is:

[0078]

[0079] Where x and y are the position coordinate information of the robot, and θ is the orientation angle of the robot at this time.

[0080] The grid information occupied is:

[0081]

[0082] Among them, R and T are rotation and translation matrices, It is the initial placeholder grid information.

[0083] (2) Annotate the geometric information and semantic information of objects in the working environment, including road length, width, road signs, parking space types, intersections, entrances and exits, etc., and form a working environment road network diagram G = (V, E, F) based on semantic elements, where V represents a set of semantic nodes, each of which maps a connected area with the same semantics in the grid map; E represents a set of edges connecting each node of the robot, representing a grid of cross-repeated areas mapped by different semantic nodes; and F represents the constraints imposed by the semantic nodes on the robot's passability. In the semantic road network diagram, the path from the starting node V0 to the target node V g The shortest path is preset. The semantic road network graph of the working environment - grid subgraph is stored in the form of a discretized database.

[0084] In addition, grids with the same semantic information are clustered, and a semantic road network is constructed at the upper layer. The semantic map nodes define the scene information and constraints, and the preset paths based on experience are stored between the nodes. The working environment road network is formed based on semantic elements. Each semantic node maps the connected area with the same semantics in the grid map; the set of connecting edges of each robot node represents the cross-repeated grids in the area mapped by different semantic nodes; the semantic node includes constraints on the robot's passability. The shortest path from the starting node to the target node is preset in the semantic road network. The semantic road network map-grid subgraph of the working environment is stored in the form of a discretized database.

[0085] (3) Classify the types of robot path conflicts, and design conflict resolution strategies such as waiting in place, driving perpendicular to the original direction, and driving in the opposite direction according to different conflict types, to form a conflict type-resolution strategy template library. Define the robot conflict type as c = [c(1), c(2), c(3), c(4)], representing fixed-point conflict, edge conflict, following conflict, and loop conflict in the robot path; assume that the robot can execute actions a = [a(1), a(2), a(3), a(4)], representing waiting in place, moving straight, moving sideways, and rotating. The execution time cost of each action is cost = [cost1, cost2, cost3, cost4] time steps. The robot conflict resolution strategy is to execute actions that are inconsistent with the original path requirements in order to resolve the conflict. The expression is:

[0086] π - =[π - (1), π - (2), π - (3)]

[0087] Among them, π - (1), π - (2), π - (3) represent waiting at the original location, driving perpendicular to the original direction, and driving in the opposite direction.

[0088] in:

[0089] π - (1) = [a(1) * w1]

[0090] π - (2) = [a(4)a(.)*w2]

[0091] π - (3) = [-a k-1 *w3]

[0092] Where w1, w2, w3 are the optimizable parameters of the selected action execution time step, a(.) represents the rotation direction action, -a k-1 Indicates the reverse action of the previous step. Figure 3 As shown, a feasible elimination strategy mapping table is established according to the conflict type.

[0093] In order to improve the accuracy of conflict type classification, the present invention combines the environmental perception module and the deep learning model (for example, the multi-layer MLP model) to realize the dynamic identification and classification adjustment of the conflict type. First, the environmental perception sensor data and the status information of multiple robots are collected as input data, and the data is cleaned, normalized and feature extracted, and the output is different types of conflict labels. Then, by inputting the historical data of multi-robot path planning and obstacle avoidance, the deep learning model is trained, and the parameters are continuously optimized to accurately classify different conflict types (vertex conflict, edge conflict, follow-up conflict). According to different conflict types, corresponding conflict resolution strategies are formulated, and a mapping table of conflict types and resolution strategies is established.

[0094] (4) When the planning system is running online, the bottom layer searches the semantic road network map to form an initial path based on the initial position and destination information of each stand-alone robot. Specifically:

[0095] When the robot is running in online planning, it receives the initial and final position information of multiple robots, and selects a preset path v0->,...,->v for each robot based on the preset route of the preset road network graph. h ->v g . Among them, -> indicates the direction of the path, v h ->v g Indicates that from v h Walk to v g For each stand-alone robot, the Hybrid A* algorithm is used to generate the shortest path for the connectable semantic nodes in the grid subgraph to form an initial path connecting the robot's initial and final positions. The Hybrid A* algorithm is a graph search algorithm.

[0096] (5) Detect and classify the paths of multiple robots, select the sub-grid map paths with spatial conflicts and pass them to the upper-level planner. Specifically:

[0097] Conflict detection is performed on the repeated semantic nodes of multiple robots. The robots perform actions at time intervals of Δt as π = [a1, ..., a K ] is a sequence of actions, and a conflict model including collision detection and potential collision detection is designed at k time steps:

[0098] It means that at time step k, robots i and j appear in the same path and occupy a node.

[0099] It means that at time step k, robot i and robot j simultaneously cross the edge and exchange nodes.

[0100] It means that robot i will occupy the current robot j node at the next moment and follow too closely.

[0101]

[0102] Indicates that multiple robots form a ring, which will cause deadlock when following conflicts are prohibited.

[0103] Among them, a K represents the final action, π i [k] represents the path node of robot i at time step k, π j [k] represents the path node of robot j at time step k, π j [k+1] represents the path node of robot j at k+1 time step, π i [k+1] represents the path node of robot i at the k+1 time step, π i+1 [k] means at k time step, π i+1 [k+1] represents the path node of robot i+1 at k+1 time step, π i+2 [k] represents the path node of robot i+2 at time step k, π j-1 [k+1] represents the path node of robot j-1 at the k+1 time step.

[0104] (6) The upper-level planner determines the priority of the robots based on the entanglement complexity of the multi-robot paths and outputs the robot conflict sub-paths with low priority. Specifically:

[0105] The robot priority is determined by the complexity of path entanglement in multiple robots. First, the number of conflicts between each path and other paths is determined. Each pair of paths is analyzed, and the number of times they occupy the same position at the same time is counted. All the conflict times are accumulated as the total conflict times. The frequency of conflicts with other paths is defined as the entanglement degree of the two paths:

[0106] Conflict(π i , π j ) = SUM k->K [C(1), C(2), C(3), C(4)]

[0107] In the formula, K represents the end point of the path, and k->K represents the process from the starting point to the end point.

[0108] Based on the priority rule of conflict resolution, since multiple robots have multiple destinations, the sum of the conflict indexes between the designed robot path and other robots in the scene is:

[0109] Score(π i ) = SUM j≠i [Conflict(π i , π j )]

[0110] It can be seen that when the path π i The more conflicts there are with other paths, the lower the priority of the robot. The robot needs to enter avoidance and replanning; in particular, when the degree of entanglement of the paths is equal, the path with longer execution time has higher priority to avoid increasing the overall running time of the planning system; when the degree of entanglement and path time are equal, the robot priority is determined to be the same and the avoidance party and the priority party are randomly selected;

[0111] (7) The upper-level planner calls the conflict type-resolution strategy template library, calculates the optimal strategy for the resolution strategy cost function, performs spatiotemporal path redistribution and parameter optimization based on the conflict resolution strategy, and outputs a conflict-free resolution path. Specifically:

[0112] By calculating the comprehensive cost of the conflict resolution solution, the path with the minimum total cost is selected. The comprehensive cost takes into account the time cost, energy cost and path deviation to ensure the high efficiency of multi-robot collaboration under the premise of ensuring safety. The spatiotemporal optimization algorithm is used to replan the path in the sub-grid map, and the time redistribution is used to optimize the low-priority robot conflict nodes in the grid sub-map. First, the initial path is planned for each robot in the sub-grid, and then the time is allocated according to the priority of the conflict node, so that the low-priority robot appropriately avoids when encountering a high-priority robot. Finally, by optimizing the path completion time of each robot, the maximum completion time at the conflict node is minimized, thereby improving the efficiency and coordination of the overall path planning. The maximum completion time of h conflicting robots to replan the conflicting path nodes:

[0113]

[0114] Among them, each The shortest time for the h-th conflicting robot to find a collision-free path in the sub-grid map. A single conflicting robot iterates through the constraint tree. Each iteration can solve a pair of conflicts according to the constraints until a conflict-free path is found. At this point, the optimization function replans the conflicting nodes in the spatiotemporal space. The original path and the replanned path are integrated to finally generate a conflict-free optimal time strategy path for multiple robots. w' represents the optimization weight parameter.

[0115] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A multi-robot hierarchical spatiotemporal optimization path planning method based on conflict resolution, characterized in that: The method includes: S1. Constructing a semantic topology and a grid hierarchical graph of a multi-robot operation environment, and using a graph search algorithm in the semantic topology and the grid hierarchical graph to generate a preset robot path from a start node to a target node; S2. Based on the pre-configured classification rules, the robot path conflict type is preliminarily detected, and the pre-trained deep learning model is used to correct the preliminary detection of the robot path conflict type to obtain the detection and classification results of the robot path conflict; S3, calculating the corresponding dynamic priority of the robot with path conflict according to the entanglement degree of the robot path and the path execution time; S4. Based on the detection and classification results of the robot path conflict, the pre-established conflict type resolution strategy template library and the cost function, the total cost of each conflict resolution strategy of the robot path conflict is evaluated; according to the principle of minimum total cost, the optimal conflict elimination strategy for each robot is obtained; S5. Filter out the robots to be processed according to their corresponding dynamic priorities; and replan the paths of the robots to be processed based on the optimal conflict elimination strategies and preset paths of the robots to be processed.

2. The multi-robot hierarchical spatiotemporal optimization path planning method based on conflict resolution according to claim 1 is characterized in that: The method of constructing a semantic topology and a grid hierarchical graph of a multi-robot operation environment, and generating a robot preset path from a starting node to a target node using a graph search algorithm in the semantic topology and the grid hierarchical graph includes: Import the working environment grid map and element information, and perform spatiotemporal gridding on the map. The robot width information and the safety distance between robots are used as the spatial resolution of the working environment grid map, and the time interval of the grid occupied is determined according to the robot driving speed. Label and cluster the semantic information of objects in the robot's working environment, and form a semantic road network diagram of the working environment based on the semantic information of objects; The semantic road network map of the working environment is merged with the grid map of the working environment to obtain the semantic topology and grid hierarchical map. In the semantic topology and grid hierarchical map, the shortest path from the start node to the target node generated based on the graph search algorithm is used as the preset path of the robot.

3. The multi-robot hierarchical spatiotemporal optimization path planning method based on conflict resolution according to claim 1 is characterized in that: Based on the pre-configured classification rules, the robot path conflict type is preliminarily detected, and the pre-trained deep learning model is used to correct the preliminary detection of the robot path conflict type, and the robot path conflict detection classification results are obtained, including: Using pre-configured classification rules, obtain preliminary detection and classification results of robot path conflict types; The environmental perception module is used to obtain the environmental perception sensor data of the multi-robot working environment, and combined with the robot's state information to obtain input data, and obtain different types of robot conflict labels as output data; Train the pre-configured multi-layer MLP model using input data and output data, and optimize the parameters of the multi-layer MLP model; Using the detection and classification results output by the multi-layer MLP model, the preliminary detection and classification results are corrected by weighted adjustment to obtain the detection and classification results of the robot path conflict.

4. The multi-robot hierarchical spatiotemporal optimization path planning method based on conflict resolution according to claim 3 is characterized in that: The method of obtaining the preliminary detection and classification results of the robot path conflict type by using the pre-configured classification rules includes: When the paths of the robots intersect, the type of robot path conflict is determined to be a fixed-point conflict; When the robot group crosses the edge and exchanges nodes at the same time, the type of robot path conflict is judged to be fixed-point conflict; When the node where the robot is located is occupied at the next moment, the type of robot path conflict is determined to be a following conflict; When the robot group forms a ring, the type of robot path conflict is judged to be a ring conflict.

5. The multi-robot hierarchical spatiotemporal optimization path planning method based on conflict resolution according to claim 1 is characterized in that: The step of calculating the dynamic priority of the robot with path conflict according to the entanglement degree of the robot path and the path execution time includes: Obtain the total number of conflicts between each robot path and other paths, and assign priorities to robots based on the principle that the greater the total number of conflicts, the lower the priority of the robot; When the total number of conflicts in the robot group is the same, the priority of the robots is assigned according to the principle that the shorter the path execution time, the lower the priority of the robot; When the total number of conflicts and execution time of the robot group are the same, the avoidance party and the priority party are randomly selected.

6. The multi-robot hierarchical spatiotemporal optimization path planning method based on conflict resolution according to claim 1 is characterized in that: The pre-established conflict type resolution strategy template library includes: According to the different conflict types of robots, formulate corresponding conflict resolution strategies; based on the conflict resolution strategies, establish a conflict type resolution strategy template library.

7. The multi-robot hierarchical spatiotemporal optimization path planning method based on conflict resolution according to claim 1 is characterized in that: The total cost of each conflict resolution strategy for the robot path conflict is evaluated based on the detection and classification results of the robot path conflict, the pre-established conflict type resolution strategy template library and the cost function, including: By defining different robot path conflict types and conflict type resolution strategies, a cost function is established; the weights in the cost function are adjusted according to the pre-set logic; Based on the detection and classification results of the robot path conflict and the pre-established conflict type resolution strategy template library, several conflict type resolution strategies are obtained; and the total cost corresponding to each conflict type resolution strategy is calculated using the cost function.

8. The multi-robot hierarchical spatiotemporal optimization path planning method based on conflict resolution according to claim 7 is characterized in that: The weights in the cost function are adjusted according to a preset logic including: When the obstacle density is higher than the preset obstacle density threshold, the safety weight is increased; When the obstacle density is lower than the preset obstacle density threshold, the path efficiency weight is increased; When the robot density on the robot path is higher than the preset robot density threshold, the avoidance and coordination weights are increased; When the robot density on the robot path is lower than the preset robot density threshold, the speed weight is increased; When an emergency is detected, the weighting of safety and response speed is increased.

9. The multi-robot hierarchical spatiotemporal optimization path planning method based on conflict resolution according to claim 1 is characterized in that: The robots to be processed are screened out according to the dynamic priorities of the robots, including: The dynamic priorities of the robots are compared, and the robot with the lowest priority is selected as the robot to be processed.

10. The multi-robot hierarchical spatiotemporal optimization path planning method based on conflict resolution according to claim 9, characterized in that: The path replanning of the robot to be processed based on the optimal conflict elimination strategy and the preset path of the robot to be processed includes: Obtain a preset path of the robot to be processed; and adjust the preset path of the robot to be processed to a replanned path by executing an optimal conflict elimination strategy of the robot to be processed.

Images