Path planning method and device, equipment and storage medium

Abstract

The invention discloses a path planning method and device, equipment and a storage medium. The method comprises the following steps: acquiring position information of a robot on a plurality of data sampling points during map construction; screening a plurality of central key points in the plurality of data sampling points based on the position information corresponding to the plurality of data sampling points; determining neighbor points of each key point in a plurality of key points obtained by expanding the central key point; and based on the position information of the multiple key points and the position information of the neighbor points of the key points, determining a moving path of the robot moving from the current position to the target position. Therefore, the working area of the robot can be effectively covered by utilizing the data sampling points in the map construction period of the robot and generating the possible track points of the robot in the map, so that a large number of invalid areas are avoided during path searching, and the efficiency of planning the robot path is improved.

Description

Technical Field

[0001] This application relates to the field of robotics, and in particular to a path planning method, apparatus, device, and storage medium. Background Technology

[0002] With the continuous evolution of intelligent manufacturing and automation technologies, mobile industrial robots, especially Automated Guided Vehicles (AGVs), are playing an increasingly prominent role in modern logistics systems. As an important carrier for realizing intelligent material transfer, AGVs demonstrate significant advantages in various aspects such as warehouse scheduling, production line collaboration, and transportation automation. Their stable and efficient operation relies on high-precision environmental modeling and positioning capabilities, while global path planning becomes the key technical support for achieving autonomous navigation and completing complex tasks.

[0003] Currently, most mainstream AGV systems employ a priori map-based positioning mechanism, utilizing LiDAR, visual recognition, or multi-sensor fusion to acquire environmental features and combining them with known map information for pose estimation. However, with the diversification of application environments and the increase in task complexity, traditional path planning methods are gradually revealing limitations in terms of adaptability and computational efficiency. Summary of the Invention

[0004] In view of the above, embodiments of this application provide at least one path planning method, apparatus, device, and storage medium.

[0005] The technical solution of this application embodiment is implemented as follows: In a first aspect, embodiments of this application provide a path planning method, the method comprising: Obtain the robot's position information at multiple data sampling points during map construction; Based on the location information corresponding to multiple data sampling points, multiple central key points are selected from the multiple data sampling points; the distance between any two central key points that are adjacent in sampling time is within a first preset distance range. Among the multiple key points obtained by expanding the central key point, the neighbor points of each key point are determined; the robot trajectory between the neighbor points and the key points does not interfere with the outline of obstacles in the constructed map; Based on the location information of multiple key points, as well as the location information of the neighboring points of each key point, the robot's movement path from its current location to its target location is determined.

[0006] Secondly, embodiments of this application provide a path planning device, the device comprising: The acquisition module is used to acquire the location information of multiple data sampling points when the robot is building the map; The filtering module is used to filter multiple central key points from multiple data sampling points based on the location information corresponding to each of the multiple data sampling points. The distance between each pair of adjacent points is within a first preset distance range. The first determining module is used to determine the neighbor points of each key point among multiple key points obtained by expanding the central key point; the robot trajectory between the neighbor points and the key points does not interfere with the outline of obstacles in the constructed map; The second determining module is used to determine the movement path of the robot from the current position to the target position based on the position information of multiple key points and the position information of the neighboring points of each key point.

[0007] Thirdly, embodiments of this application provide a computer device, including a memory and a processor. The memory stores a computer program that can run on the processor, and the processor executes the program to implement some or all of the steps in the above-described method.

[0008] Fourthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements some or all of the steps in the above-described method.

[0009] Fifthly, embodiments of this application provide a computer program product, including a computer program or instructions, which, when executed by a processor, implement some or all of the steps in the above-described method.

[0010] In this embodiment, multiple key points are generated by utilizing the center key points where the distance between any two adjacent points is within a first preset distance range during the robot's map construction. Then, based on the positional information of these key points, the robot's trajectory between each key point and other key points is determined, along with neighboring points that do not interfere with the outlines of obstacles in the constructed map. Finally, based on the multiple key points and their neighboring points, the robot's movement path from its current position to its target position is determined. This method utilizes data sampling points from the robot's map construction to generate possible trajectory points (i.e., key points) within the constructed map. This effectively covers the robot's working area, avoiding numerous invalid areas during path search and improving the efficiency of robot path planning. Simultaneously, by determining the neighboring points of each key point, reachable paths for the robot within the constructed map can be determined, thereby improving the accuracy of robot path planning.

[0011] It should be understood that the above general description and the following detailed description are merely exemplary and explanatory, and are not intended to limit the technical solutions of this application. Attached Figure Description

[0012] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with this application and, together with the specification, serve to explain the technical solutions of this application.

[0013] Figure 1 A schematic diagram of the implementation process of a path planning method provided in this application embodiment. Figure 1 ; Figure 2 A schematic diagram illustrating the screening of central key points provided in this application embodiment; Figure 3 A schematic diagram of the implementation process of a path planning method provided in this application embodiment. Figure 2 ; Figure 4 A schematic diagram illustrating the construction of subpaths provided in an embodiment of this application; Figure 5 A schematic diagram illustrating the planning of a movement path provided in an embodiment of this application; Figure 6 A schematic diagram of the implementation process of a path planning method provided in this application embodiment. Figure 3 ; Figure 7 Extended schematic diagram of candidate key points provided in the embodiments of this application; Figure 8 A schematic diagram of the implementation process of a path planning method provided in this application embodiment. Figure 4 ; Figure 9 A schematic diagram illustrating the screening of target key points provided in this application embodiment; Figure 10 This is a schematic diagram of the structure of the path planning system provided in the embodiments of this application; Figure 11 This is a schematic diagram of the path planning device provided in the embodiments of this application; Figure 12 This is a schematic diagram of a hardware entity of a computer device provided in an embodiment of this application. Detailed Implementation

[0014] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application are further described in detail below with reference to the accompanying drawings and embodiments. The described embodiments should not be regarded as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0015] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0016] The terms “first / second / third” are used merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that “first / second / third” may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0017] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. The terminology used herein is for descriptive purposes only and is not intended to limit the scope of this application.

[0018] The core objective of global path planning in related technologies is to generate a feasible path from the starting point to the target point for an AGV within the overall environment. This not only requires the path length to be as short as possible but also necessitates comprehensive consideration of multiple performance indicators such as running time, energy consumption, and safety. For industrial robots, optimizing the path structure helps improve operating efficiency, reduce energy consumption, and extend equipment lifespan. Especially in high-density work areas or environments with many moving obstacles, quickly generating a safe, efficient, and feasible path is a crucial prerequisite for autonomous robot navigation.

[0019] To address the technical problem in related technologies where paths cannot be quickly generated in scenarios with dynamic adjustments and frequent changes in obstacles, embodiments of this application provide a path planning method, such as... Figure 1 As shown, the path planning method includes steps S101 to S104: Step S101: Obtain the robot's position information at multiple data sampling points during map construction.

[0020] Here, data sampling points refer to the sampling points where environmental data is collected during the map construction process. It's understandable that to build a map of an unfamiliar environment, the robot needs to scan environmental data using its sensors, and then construct the map based on that data. Therefore, data sampling points are the sampling points where the robot collects environmental data using its sensors.

[0021] For example, when the robot's sensors are LiDAR and / or vision sensors, data sampling points can refer to sampling points formed by the robot's LiDAR or camera acquiring information about the robot's surrounding environment. These data sampling points typically contain the robot's position information in the environment, used to describe the robot's trajectory. For instance, when the robot is operating in the environment, its own LiDAR or vision sensors will collect information about the robot's current position at each moment, thus forming a series of discrete points, which are the data sampling points. Data sampling points not only reflect the robot's own pose but also form the basis for constructing an environmental map.

[0022] Step S102: Based on the location information corresponding to the multiple data sampling points, select multiple central key points from the multiple data sampling points; the distance between any two central key points that are adjacent in sampling time among the multiple data sampling points is within a first preset distance range.

[0023] In this embodiment, the distance between any two adjacent central key points among multiple data sampling points is within a first preset distance range. For example, this first preset distance range can be 0.5 meters to 1 meter. It is understood that to improve the accuracy of robot mapping, the robot needs to collect a large amount of environmental data; that is, the robot collects data at a high frequency, resulting in a large number of data sampling points. To improve data processing efficiency, multiple central key points that meet the conditions can be filtered from multiple data sampling points based on the first preset distance range, thereby reducing unnecessary data volume and improving path generation efficiency.

[0024] In this embodiment, multiple central key points can be sequentially selected from the data sampling points according to their acquisition time sequence and location information, with the distance between each pair of adjacent points within a first preset distance range. Finally, connecting the adjacent central key points yields a path composed of multiple central key points throughout the robot's mapping process. The starting central key point of this path is the robot's mapping start point, and the ending central key point is the robot's mapping end point.

[0025] For example, such as Figure 2 As shown, the robot collects environmental data along the mapping trajectory, thereby constructing a map 201 based on the environmental data. The mapping trajectory includes multiple data sampling points for collecting environmental data, and a central key point 2021 is selected from the multiple data sampling points by means of a first preset distance range.

[0026] Step S103: Among the multiple key points obtained by expanding the central key point, determine the neighbor points of each key point; the robot trajectory between the neighbor points and the key points does not interfere with the outline of the obstacles in the constructed map.

[0027] In this embodiment, based on the location information of each central key point, an extended region can be formed by offsetting a preset distance in a preset direction along the mapping trajectory, and candidate points can be uniformly acquired within this extended region. For example, the preset direction can be the normal direction of the mapping trajectory.

[0028] In some embodiments, after expanding each central keypoint, the expanded keypoints can be downsampled to filter the multiple candidate points. The filtered candidate points and the central keypoints are then used together as keypoints. This not only preserves the robot's positional features during mapping but also reduces the number of candidate points, thereby improving path planning efficiency.

[0029] Understandably, the purpose of keypoint expansion is to enrich the keypoints, allowing them to better adapt to path planning needs in complex environments. Compared to methods that only use the central keypoint for path planning, expanded keypoints provide more feasible path options, reducing the probability of path search failure. For example, in narrow passages or environments with local obstacles, relying solely on the central keypoint may result in an unconnected path, while expanded keypoints may allow for bypassing obstacles and finding an effective path.

[0030] In this embodiment, after obtaining multiple key points by expanding the central key point, it is necessary to filter the neighboring points of the current key point among the multiple key points. The filtering criterion for the neighboring points of a key point is that the robot trajectory between the key point and the current key point does not interfere with the outlines of obstacles in the constructed map. In some embodiments, the filtering criterion for the neighboring points of a key point may further include that the distance between the neighboring points and the current key point is within a fourth preset distance range. For example, this fourth preset distance range may be 0.8 meters to 1 meter.

[0031] In this embodiment, multiple candidate key points that are within a fourth preset distance range from the current key point are first determined from multiple other key points. Then, the width information of the robot's projection on the map is obtained. Based on the width information, the robot trajectory is constructed on the line connecting the key point and other candidate key points. If the robot trajectory does not interfere with the outline of the obstacle in the constructed map, the other candidate key point is determined as a neighbor point of the current key point.

[0032] Step S104: Based on the location information of multiple key points and the location information of the neighboring points of each key point, determine the movement path of the robot from the current position to the target position.

[0033] Here, the robot's current position can refer to the robot's location within the constructed map, while the target position is the location the robot wants to reach within the constructed map.

[0034] In this embodiment, based on the location information of multiple key points, the robot's current position, and the robot's target position, the starting key point closest to the current position and the ending key point closest to the target position are determined from among the multiple key points. Then, based on the location information of the starting key point, the ending key point, and the location information of the neighboring points of each key point, multiple target key points with the shortest movement time from the starting key point to the ending key point are selected from among multiple other key points. The trajectory corresponding to the starting key point, multiple target key points, and the ending key point is determined as the movement path of the robot from the current position to the target position. Among these, other key points are key points other than the starting and ending key points, and target key points are the neighboring points of the key points. That is, it is necessary to ensure that the key points on the finally determined movement path are the neighboring points of other key points. This ensures that the determined movement path is reachable by the robot and will not interfere with obstacles in the constructed map.

[0035] Understandably, path planning methods in related technologies are generally probabilistic graph-based stochastic path planning methods. The drawback of these methods is the need to determine the generation method of random points and the route map between them. If too few random points are selected, path planning may fail; if too many are selected, the solution time may increase exponentially, compromising real-time performance. In this embodiment, data sampling points during the robot's map construction are used to generate possible trajectory points (i.e., key points) within the constructed map. This effectively covers the robot's working area, avoiding a large number of invalid areas during path search and significantly improving search speed.

[0036] In this embodiment, multiple key points are generated by utilizing the center key points where the distance between any two adjacent points is within a first preset distance range during the robot's map construction. Then, based on the positional information of these key points, the robot's trajectory between each key point and other key points is determined, along with neighboring points that do not interfere with the outlines of obstacles in the constructed map. Finally, based on the multiple key points and their neighboring points, the robot's movement path from its current position to its target position is determined. This method utilizes data sampling points from the robot's map construction to generate possible trajectory points (i.e., key points) within the constructed map. This effectively covers the robot's working area, avoiding numerous invalid areas during path search and improving the efficiency of robot path planning. Simultaneously, by determining the neighboring points of each key point, reachable paths for the robot within the constructed map can be determined, thereby improving the accuracy of robot path planning.

[0037] In some embodiments, such as Figure 3 As shown, step S104 above can be achieved through steps S301 and S302: Step S301: For each of the multiple key points, based on the location information of the key point and the location information of the neighboring points corresponding to the key point, determine the sub-path between the key point and each neighboring point.

[0038] In this embodiment of the application, for each key point, the key point can be connected with its corresponding neighboring points to obtain the sub-path between the key point and each neighboring point.

[0039] For example, such as Figure 4 As shown, for the first keypoint 401, there are 3 neighboring points, namely the first neighboring point 402, the second neighboring point 403, and the third neighboring point 404. For the second keypoint (i.e., the first neighboring point 402), there are 2 neighboring points, namely the third neighboring point 404 and the fourth neighboring point 405. Connecting the first keypoint 401 with the first neighboring point 402, the second neighboring point 403, and the third neighboring point 404 respectively, we obtain three sub-paths between the first keypoint 401 and the first neighboring point 402, the second neighboring point 403, and the third neighboring point 404. Similarly, connecting the second keypoint (i.e., the first neighboring point 402) with the third neighboring point 404 and the fourth neighboring point 405 respectively, we obtain two sub-paths between the second keypoint and the third neighboring point 404 and the fourth neighboring point 405 respectively.

[0040] Step S302: Based on multiple sub-paths, determine the movement path for the robot to move from the current position to the target position.

[0041] In this embodiment, a preset path search algorithm can be used to filter target sub-paths from multiple sub-paths, and then a movement path for the robot to move from the current position to the target position can be constructed based on the multiple target sub-paths. For example, the preset path search algorithm includes at least one of the following: a neural network-based path search method and a greedy algorithm.

[0042] In this embodiment, a complete path network can be formed by finding neighboring points for each key point and establishing sub-paths. Since the selection of neighboring points takes into account interference with obstacle contours, the feasibility of the selected path is improved, thereby reducing the risk of path conflicts and enhancing the practicality of path planning.

[0043] In some embodiments, step S302 can be implemented by steps S3021 to S3023: Step S3021: Determine the key point closest to the robot's current position as the starting key point, and determine the key point closest to the robot's target position as the ending key point.

[0044] In this embodiment of the application, based on the location information of each key point, as well as the robot's current position and target position, the starting key point closest to the robot's current position and the ending key point closest to the robot's target position can be determined among multiple key points.

[0045] For example, such as Figure 5 As shown, among multiple key points, the starting key point 502, which is closest to the robot's current position 501, and the ending key point 504, which is closest to the robot's target position 503, are determined.

[0046] Step S3022: Based on the location information of the starting key point and the location information of the ending key point, filter the target sub-path from multiple sub-paths.

[0047] In this embodiment of the application, the A* search algorithm can be used to filter path key points from multiple key points based on the location information of the starting key point and the location information of the ending key point. Then, the sub-path between the multiple path key points, the starting key point and the ending key point is determined as the target sub-path.

[0048] In this embodiment of the application, step S3022 can be implemented through the following steps: Step 1. Add the neighboring points of the starting key point to the open list; Step 2. Determine the neighbor with the shortest cost among the neighboring points of the starting key point using the cost function, and add the neighbor with the shortest cost to the closed list; wherein, the above cost function can be implemented by formula (1): (1); Where g(n) represents the Euclidean distance cost from the starting keypoint to the current neighbor, and h(n) represents the estimated Euclidean distance cost between the current neighbor and the ending keypoint.

[0049] Step 3. Add the neighboring point with the shortest distance cost to the closed list in turn, until the open list is empty or the current neighboring point is the termination key point.

[0050] If the open list is empty, it means the path search failed. If the current neighbor point is the termination key point, the path between each neighbor point in the list is closed, representing the shortest path between the start key point and the termination key point.

[0051] For example, such as Figure 5 As shown, the A* search algorithm filters out the first path key point 505, the second path key point 506, and the third path key point 507 from multiple key points. Then, the sub-path between the starting key point 502 and the first path key point 505, the sub-path between the second path key point 506 and the third path key point 507, and the sub-path between the third path key point 507 and the ending key point 504 are determined as the target sub-paths.

[0052] Step S3023: Based on the starting path between the robot's current position and the starting key point, the ending path between the robot's target position and the ending key point, and multiple target sub-paths, construct the robot's movement path.

[0053] In this embodiment, the robot's current position and starting key point can be connected to obtain the starting path, and then the robot's target position and ending key point can be connected to obtain the ending path. Finally, the starting path, ending path, and multiple target sub-paths are connected to obtain the robot's movement path.

[0054] For example, such as Figure 5 As shown, Figure 5 The dashed line in the diagram represents the robot's movement path from its current position to the target position.

[0055] In this embodiment, based on the location information of the starting key point closest to the robot's current position, the location information of the ending key point closest to the robot's target position, the robot's current position, and the robot's target position, a target sub-path is selected from multiple sub-paths. Based on these multiple target sub-paths, the robot's movement path is constructed. Since the key points all contain the robot's positional features during mapping, the selected target sub-paths are paths that the robot has verified and can traverse during mapping. Therefore, constructing the robot's movement path from its current position to its target position using verified target sub-paths not only improves the accuracy of path planning but also reduces the amount of data processing required, thereby improving the efficiency of path planning.

[0056] In some embodiments, such as Figure 6 As shown, the above method can also be implemented through steps S601 to S603: Step S601: For each central key point, based on the location information of the central key point, determine multiple candidate key points whose distance from the central key point is within a second preset distance range.

[0057] Here, candidate keypoints refer to potential location points selected around the central keypoint within a second preset distance range, used for subsequent filtering and path construction. For example, this second preset distance range can be from 0.4 meters to 0.6 meters.

[0058] Understandably, a mapping trajectory including multiple central key points refers to the trajectory the robot moves along while building the map. This trajectory represents the robot's effective working path within the constructed map. In other words, there are no obstacles on this trajectory, so it can be used to plan the robot's path from its current position to a target position. However, because the robot's sensors have a limited data scanning range, the robot's mapping trajectory has certain limitations. Therefore, it is necessary to expand upon the central key points to ensure that the expanded key points are evenly distributed within the robot's effective working area. This allows for the generation of multiple possible expansion points near each central key point, enriching the key points and improving coverage and success rate during path planning.

[0059] Step S602: Based on the location information of each central key point, multiple candidate key points corresponding to each central key point are filtered to obtain multiple target key points.

[0060] In this embodiment, it can be determined whether a central key point exists around each candidate key point, and / or whether a candidate key point has already been identified as a target key point. It is understood that if a central key point exists around a candidate key point, and / or whether a candidate key point has already been identified as a target key point, it indicates that there are a large number of key points in that area. In order to improve the efficiency of trajectory planning, this candidate key point can be deleted.

[0061] In some embodiments, the coordinate information of the outline of obstacles in the constructed map can be obtained. If, based on the position information of the candidate key point and the coordinate information of the outline of the obstacle, it is determined that the candidate key point is within the outline of the obstacle, it can be indicated that the candidate key point is on the obstacle, and the candidate key point can be deleted.

[0062] In this embodiment of the application, by screening candidate key points, redundant or invalid candidate points can be removed, reducing unnecessary processing burden, while improving the rationality of the distribution of key points, which helps to improve path search efficiency.

[0063] Step S603: Determine multiple target key points and multiple center key points as multiple key points, and obtain the location information of multiple key points.

[0064] In this embodiment, the multiple key points include the initial central key point of the robot during mapping, and target key points obtained by expanding the central key point. In this way, the multiple key points not only preserve the robot's positional features during mapping, but also enrich the trajectory points within the robot's effective working area during mapping.

[0065] In this embodiment, candidate key points are obtained by expanding the central key point, then the candidate key points are filtered, and finally the target key point and the central key point are combined as multiple key points. This not only preserves the robot's positional features during mapping but also enriches the trajectory points within the robot's effective working area during mapping, thereby improving the accuracy of the planned trajectory.

[0066] In some embodiments, step S601 can be implemented by steps S6011 and S6012: Step S6011: Construct a circular region with the location information of the central key point as the center and the radius within a second preset distance range.

[0067] Here, the circular region refers to a circular search area constructed around the central key point. The radius of this circular region is preset, and the radius determines the distribution density and coverage of the sampling points. For example, the second preset distance range can be between 0.4 meters and 0.6 meters.

[0068] Step S6012: Using a preset point selection method, obtain multiple candidate key points on the circular area corresponding to the central key point.

[0069] In this embodiment, to improve the uniformity of candidate keypoint distribution within the robot's effective working area, the same point-sampling method can be used to select points on the circular region corresponding to each central keypoint. Specifically, this point-sampling method can involve uniformly selecting points on the circular region, with equal distances between adjacent candidate keypoints on each circular region.

[0070] In some embodiments, the N vertices of the inscribed regular N-gon of the circular region of each central key point can be used as candidate key points corresponding to the central key point, where N is an integer greater than or equal to 3. For example, as shown... Figure 7 As shown, the six vertices of the inscribed regular hexagon of the circular region (not shown in the figure) of each central key point 2021 are taken as candidate key points 701 corresponding to the central key point 2021. Among them, the orientation of each vertex of the inscribed regular hexagon corresponding to each central key point is the same.

[0071] It is understandable that using the same point selection method for different central key points can ensure that the generation method of candidate key points in all circular areas remains consistent. This consistency can lead to a consistent layout of key points throughout the entire working area, thereby improving the coherence and stability of path search.

[0072] In some embodiments, different point-sampling methods may be used to select points on the circular region corresponding to each central keypoint. These different point-sampling methods may refer to: different numbers of candidate keypoints being obtained on the circular region corresponding to each central keypoint; the same number of points being obtained on the circular region corresponding to each central keypoint; and each vertex of the regular polygon formed by multiple candidate keypoints corresponding to each central keypoint having a different orientation.

[0073] In this embodiment, a circular region with a radius within a second preset distance range and a central key point as its center is constructed. Multiple candidate key points are then acquired within the circular region corresponding to each central key point using a preset point-sampling method. This ensures that the candidate key points are evenly distributed across the robot's effective working area, thereby improving the efficiency and success rate of path planning.

[0074] In some embodiments, such as Figure 8 As shown, step S602 above can be achieved through steps S801 and S802: Step S801: According to the sampling time sequence of the robot at multiple central key points, the screening results of multiple candidate key points for each central key point are determined sequentially based on the location information of each central key point.

[0075] Step S802: The candidate key points corresponding to the screening results that meet the screening conditions are determined as target key points.

[0076] The screening criteria include: within a third preset distance range of the candidate key points, there is no central key point, or there are no candidate key points already identified as target key points and the central key point; and The candidate keypoints are not within the outline of obstacles in the constructed map.

[0077] In this embodiment, multiple candidate keypoints from the initial central keypoint can be screened. Specifically, for the first candidate keypoint screened from the initial central keypoint, it needs to be determined whether a central keypoint exists within a third preset distance range of this candidate keypoint, and whether the candidate keypoint lies within the outline of obstacles in the constructed map. This is because, since this candidate keypoint is the first one screened from the initial central keypoint, there are no candidate keypoints already identified as target keypoints. Therefore, it is necessary to determine whether a central keypoint exists within the third preset distance range of this candidate keypoint, and whether the candidate keypoint lies within the outline of obstacles in the constructed map.

[0078] For the second candidate keypoint among the initial center keypoints, if the first candidate keypoint is determined as the target keypoint, then the second candidate keypoint needs to be determined whether there exists a center keypoint or a candidate keypoint already determined as the target keypoint within a third preset distance range of this candidate keypoint, and whether this candidate keypoint is not within the outline of obstacles in the constructed map. After determining the screening results of multiple candidate keypoints of the initial center keypoint in sequence, multiple candidate keypoints of center keypoints adjacent to the initial center keypoint in sampling time are screened in sequence until all candidate keypoints of all center keypoints have been screened.

[0079] For example, such as Figure 9 As shown, for the initial center keypoint 901, the first candidate keypoint 9011 is selected as the target keypoint because there are no other candidate keypoints already identified as target keypoints, there are no center keypoints within the third preset distance range of the first candidate keypoint 9011, and the first candidate keypoint 9011 is not within the outline of obstacles in the constructed map. Figure 9As shown, all six candidate keypoints of the initial center keypoint 901 are target keypoints. Then, the selection results for the candidate keypoints of the next center keypoint 902 are determined. Because the second candidate keypoint 9021 and the third candidate keypoint 9022 of center keypoint 902 are within the outline of the obstacle, they are not considered target keypoints. Similarly, the fourth candidate keypoint 9023 and the fifth candidate keypoint 9024, because there are already candidate keypoints identified as target keypoints within their own third preset distance range, are also not considered target keypoints. Therefore, of the six candidate keypoints of center keypoint 902, two are target keypoints. Figure 9 As shown, after filtering the candidate key points for each central key point, multiple target key points marked as hollow circles can be obtained.

[0080] In this embodiment, the method of screening candidate key points based on sampling time order and combining it with an obstacle detection mechanism can effectively solve the problem of key point redundancy or conflict, thereby optimizing the path search efficiency and further improving the stability and reliability of global path planning.

[0081] In some embodiments, step S603 can also be implemented by steps S6031 and S6032: Step S6031: Take the target key point among multiple key points as the center key point, and repeat the steps of determining and filtering candidate key points until the number of iterations is within the preset range, and obtain multiple target key points after iteration.

[0082] Step S6032: Determine multiple key points as multiple key points after multiple loops and multiple center key points, and obtain the location information of multiple key points.

[0083] In this embodiment, candidate key points already identified as target key points can undergo further key point expansion and filtering steps, and this process can be repeated multiple times until the number of iterations is within a preset range, resulting in multiple target key points after iteration. For example, if the constructed map is an indoor map, the preset number of iterations can be 2-3 times; if the constructed map is an outdoor map, the preset number of iterations can be 4-5 times.

[0084] In this embodiment of the application, the key point expansion method for candidate key points that have been determined as target key points can be the same as the key point expansion method for the first time for the center key point. For example, the six vertices of the inscribed regular hexagon of the circular region of each center key point (or candidate key points that have been determined as target key points) can be used as candidate key points corresponding to the center key point. The key point expansion method can also be different. For example, for the center key point, the six vertices of the inscribed regular hexagon of the circular region of the center key point can be used as candidate key points corresponding to the center key point. For candidate key points that have been determined as target key points, the eight vertices of the inscribed regular octagon of the circular region of the candidate key points that have been determined as target key points can be used as candidate key points corresponding to the center key point.

[0085] In some embodiments, after step S603, the above method may further include: downsampling multiple key points to obtain multiple processed key points.

[0086] In this embodiment, voxel filtering can be used to downsample multiple key points.

[0087] In some embodiments, step S6031 may include: taking the target key point among multiple processed key points as the center key point, and repeatedly performing the candidate key point determination and screening steps and the downsampling step until the number of iterations is within a preset range, thereby obtaining multiple target key points after iteration.

[0088] In some embodiments, the above method may also be implemented through the following steps: Step S11: Obtain environmental data corresponding to multiple data sampling points, as well as the robot's initial posture.

[0089] Step S12: Based on the initial environmental data of the robot in its initial posture and the adjacent environmental data of adjacent sampling points, determine the robot's relative spatial transformation.

[0090] Adjacent sampling points refer to data sampling points that are adjacent to the data sampling points corresponding to the initial environmental data in terms of sampling time.

[0091] Step S13: Based on the robot's relative spatial transformation and the robot's initial posture, determine the robot's adjacent postures at adjacent sampling points.

[0092] In this embodiment of the application, the robot's adjacent poses can be achieved using formula (2): Formula (2); in, T represents the robot's relative spatial transformation, T1 represents the robot's initial pose, and T2 represents the robot's adjacent pose.

[0093] Step S14: Repeat the above steps to obtain the robot's posture information at each data sampling point.

[0094] Step S15: Based on the robot's posture information at each data sampling point and the environmental data corresponding to multiple data sampling points, construct a 3D map.

[0095] Step S16: Select location points within the height range between the ground and the robot's height on the 3D map, and project multiple location points onto a 2D plane to obtain a 2D map.

[0096] In this embodiment of the application, after processing steps S11 to S16, the final result is... Figure 2 Map 201.

[0097] This application provides a path planning method that can be applied to automated guided vehicles (AGVs). The path planning method can be implemented through the following steps: Step S21: Receive the task instruction sent by the host computer; the task instruction carries the target location of the automated guided vehicle.

[0098] The host computer can be a Manufacturing Execution System (MES) or a Warehouse Management System (WMS).

[0099] Step S22: In response to the task instruction, obtain the location information of the indoor automated guided vehicle at multiple data sampling points during map construction.

[0100] Step S23: Based on the location information corresponding to the multiple data sampling points, select multiple central key points from the multiple data sampling points.

[0101] Step S24: Among the multiple key points obtained by expanding the central key point, determine the neighbor points of each key point.

[0102] Step S25: Based on the location information of multiple key points and the location information of the neighboring points of each key point, determine the movement path of the automated guided vehicle from the current location to the target location.

[0103] Step S26: Move to the target location according to the movement path and send feedback information to the host computer to indicate that the automated guided vehicle has reached the target location.

[0104] The following describes the application of the path planning method provided in this application in a real-world scenario: With the continuous evolution of intelligent manufacturing and automation technologies, mobile industrial robots, especially automated guided vehicles (AGVs), are playing an increasingly prominent role in modern logistics systems. As an important carrier for realizing intelligent material transfer, AGVs demonstrate significant advantages in various aspects such as warehouse scheduling, production line collaboration, and transportation automation. Their stable and efficient operation relies on high-precision environmental modeling and positioning capabilities, while global path planning has become a key technological support for achieving autonomous navigation and completing complex tasks.

[0105] Currently, most mainstream AGV systems employ a prior map-based positioning mechanism, utilizing LiDAR, visual recognition, or multi-sensor fusion to acquire environmental features and combine this with known map information for pose estimation. However, with the diversification of application environments and the increase in task complexity, traditional path planning methods are gradually revealing limitations in terms of adaptability and computational efficiency. Especially when facing scenarios with dynamic path adjustments and frequent obstacle changes, relying solely on local obstacle avoidance strategies is insufficient to ensure the successful completion of the task.

[0106] The core objective of global path planning is to generate a feasible path from the starting point to the target point for an AGV within the overall environment. This requires not only minimizing the path length but also comprehensively considering multiple performance indicators such as running time, energy consumption, and safety. For industrial robots, optimizing the path structure helps improve operating efficiency, reduce energy consumption, and extend equipment lifespan. Especially in high-density work areas or environments with many moving obstacles, quickly generating a safe, efficient, and feasible path is a crucial prerequisite for autonomous robot navigation.

[0107] Currently, the tasks performed by robots are relatively simple, mostly point-to-point navigation requirements. This invention recognizes that the robot's working area and mapping area highly overlap, allowing the mapping trajectory to be used as a reference route for the navigation path. Based on this background, this invention proposes a fast point-to-point global path planning method that can quickly and efficiently generate a global navigation path from the starting point to the destination.

[0108] Probabilistic graph-based stochastic path planning methods in related technologies have a drawback: they require confirmation of the generation method of random points and the route map between them. If too few random points are selected, path planning may fail; if too many are selected, the solution time may increase exponentially, compromising real-time performance. Furthermore, neither of these methods fully utilizes known robot paths; they are entirely random. Fully utilizing manually mapped trajectories or other known robot reference trajectories could further improve the accuracy and reliability of the methods.

[0109] This application provides a robot path planning method based on probabilistic route maps. Compared with existing probabilistic route planning methods, this application selects sampling points based on certain rules rather than being completely random. This strategy can effectively balance success rate and efficiency, and solve the sampling point generation problem of existing methods.

[0110] This application uses the path during mapping to generate sampling points, which can effectively cover the robot's working area. This strategy can effectively solve the problem of path solving failure caused by the randomness of sampling points in existing methods.

[0111] like Figure 10 As shown, the path planning system includes a mapping module 1001, a sampling point generation module 1002, a downsampling module 1003, a keyframe selection module 1005, a neighbor point generation module 1006, and a path search module 1007. The mapping module 1001 is used to: acquire environmental information about the robot’s surroundings as perceived by the mapping sensor 1004 installed on the robot; and construct a two-dimensional (2D) grid map by registering continuous environmental information.

[0112] The keyframe selection module 1005 is used to receive the robot pose at different times from the mapping module and select the pose corresponding to the keyframe by means of a distance threshold.

[0113] The sampling point generation module 1002 is used to add the points where the key frame pose is located to the sampling point set and expand the sampling points in the sampling point set.

[0114] The downsampling module 1003 is used to downsample all sampling points in the sampling point set using voxel filtering, and then pass the downsampled sampling point set to the sampling point generation module to obtain a new sampling point set.

[0115] The neighbor point generation module 1006 is used to sequentially find the neighbor points around each sampling point.

[0116] The path search module 1007 is used to search for an optimal path 1008 from the starting point to the ending point from the set of sampling points.

[0117] In this embodiment, assuming two adjacent times t1 and t2, the robot's initial pose at time t1 is T1. The robot's relative spatial transformation is obtained as ΔT by registering the laser point clouds or images at times t1 and t2. Therefore, the robot's pose at time t2 is: T2 = T•T1, where T represents a transformation matrix containing 3 degrees of freedom of rotation and 3 degrees of freedom of translation. The mapping module obtains the robot pose at different times by continuously registering continuous laser point clouds or images, and uses the pose at different times to stitch sensor data to construct a 3D point cloud map. Then, a point cloud map of a certain height is cropped to generate a 2D raster map.

[0118] In some embodiments, the conversion of a 3D point cloud map to a 2D grid map typically involves cropping points within a height range between the ground and the robot's height. These 3D points are then projected onto a 2D plane to obtain a 2D grid map. The 2D grid map is a grayscale image where white areas represent drivable areas for the robot and black areas represent areas where obstacles prevent the robot from traveling. For example, the 2D grid map constructed by the mapping module looks like this: Figure 2 The map shown is 201.

[0119] In this embodiment, the distance threshold for selecting keyframe poses is set to 0.8 meters. When the mapping module starts (time t0), the robot pose T0 at time t0 is selected as a keyframe. If the translational change between the pose T1 at time t0 and time t1 exceeds the distance threshold, then T1 at time t1 is determined to be a keyframe. Otherwise, the pose T2 at time t0 and time t2 are compared until a new keyframe pose is selected. Once a new keyframe is selected, its pose is saved, and the latest keyframe pose is updated for selecting the next keyframe. Ultimately, during the entire mapping process, a path composed of discrete keyframe poses is obtained, i.e. Figure 2 The mapping trajectory in the middle is 202.

[0120] In this embodiment, the sampling point generation module sequentially and uniformly selects six candidate points from the circumcircle of a regular hexagon centered on the pose of each keyframe, using the same point selection method. The radius of the circumcircle is set to 0.5 meters. The sampling point generation module sequentially determines whether there are sampling points around these six candidate points. For a point on the regular hexagon, if there are no sampling points within 0.5 meters of the point and the point is not within an obstacle, the point is added to the sampling point set; otherwise, the point is ignored. The sampling point set obtained by the sampling point generation module may include... Figure 6 Multiple target key points marked as hollow circles, and center key points marked as solid circles (i.e., keyframe pose).

[0121] In this embodiment, the voxel side length of the voxel filter used for downsampling by the downsampling module is 0.5 meters. In one loop, the sampling point generation module generates more sampling points using the sampling point set (for the first loop, the sampling point set is the location points of each keyframe), and then downsamples the new sampling point set. This loop process is performed three times in the method of this invention to ensure that there are uniformly distributed sampling points within the robot's working area.

[0122] In this embodiment of the application, the neighbor point generation module sequentially searches for all sampling points within a 1-meter (or alternatively, 0.9m) radius around each sampling point. If the robot moves along a straight line between two sampling points without colliding with obstacles on the map, then the found sampling point is considered to be a neighbor point of the current sampling point. Then, the sampling point is connected to each of its own neighbor points.

[0123] In this embodiment, the path search module uses the A* algorithm to search for an optimal path from the starting point to the ending point from the set of sampling points. The path search module first finds the sampling point s closest to the starting point and the sampling point e closest to the ending point in the set of all sampling points. Then, starting from sampling point s, the A* search algorithm is used to find the shortest path between sampling point s and sampling point e. The path search module first adds the neighboring points of sampling point s to the open list; then, it selects the neighboring point with the shortest Euclidean distance cost through the cost function and adds it to the closed list. Finally, the path search module sequentially selects the neighboring point with the lowest cost and adds it to the closed list until the open list is empty or the current neighboring point is sampling point e. If the open list is empty, it means that the path search has failed. If the current neighboring point is sampling point e, it means that a path has been found. The shortest path between sampling point s and sampling point e is obtained by backtracking the closed list. The cost function can be expressed by the above formula (1).

[0124] In this embodiment, the path search module is the output module of the path planning system. It sequentially connects the start point and the sampling point s and the sampling points searched by the path search module between sampling point s and sampling point e, and finally connects to the end navigation destination to obtain an optimal global path from the start point to the end point.

[0125] In this embodiment, the robot's working area is obtained using a mapping trajectory, and sampling points are selected within the working area based on a rule-based method. This results in a more even distribution of sampling points within the robot's working area compared to random sampling, avoiding numerous invalid regions during path search and significantly improving search speed. Furthermore, for narrow areas, because this embodiment references the mapping trajectory, a global path can be found using it, significantly improving the path planning success rate compared to methods without reference trajectories. In addition, this embodiment generates sampling points evenly within the robot's mapping area (working area), achieving a balance between path search speed and success rate compared to randomly generating sampling points, reducing path search failures and excessively long search times.

[0126] Figure 11 This is a schematic diagram of the composition structure of a path planning device provided in an embodiment of this application, as shown below. Figure 11 As shown, the path planning device 1100 includes: an acquisition module 1101, a filtering module 1102, a first determination module 1103, and a second determination module 1104; wherein, The acquisition module 1101 is used to acquire the location information of multiple data sampling points when the robot is building the map; The filtering module 1102 is used to filter multiple central key points from multiple data sampling points based on the location information corresponding to each of the multiple data sampling points; the distance between any two central key points that are adjacent in sampling time among the multiple data sampling points is within a first preset distance range. The first determining module 1103 is used to determine the neighbor points of each key point among multiple key points obtained by expanding the central key point; the robot trajectory between the neighbor points and the key points does not interfere with the outline of obstacles in the constructed map; The second determining module 1104 is used to determine the movement path of the robot from the current position to the target position based on the position information of multiple key points and the position information of the neighboring points of each key point.

[0127] In some embodiments, the second determining module 1104 is further configured to, for each of the multiple key points, determine a sub-path between the key point and each neighboring point based on the location information of the key point and the location information of the neighboring points corresponding to the key point; and determine a movement path for the robot to move from the current position to the target position based on the multiple sub-paths.

[0128] In some embodiments, the second determining module 1104 is further configured to determine the key point closest to the robot's current position as the starting key point and the key point closest to the robot's target position as the ending key point; filter target sub-paths among multiple sub-paths based on the position information of the starting key point and the position information of the ending key point; and construct the robot's movement path based on the starting path between the robot's current position and the starting key point, the ending path between the robot's target position and the ending key point, and the multiple target sub-paths.

[0129] In some embodiments, the path planning device 1100 further includes an extension module, which is further configured to, for each central key point, determine multiple candidate key points whose distance from the central key point is within a second preset distance range based on the location information of the central key point; filter the multiple candidate key points corresponding to each central key point based on the location information of each central key point to obtain multiple target key points; and determine the multiple target key points and the multiple central key points as multiple key points to obtain the location information of the multiple key points.

[0130] In some embodiments, the extension module is further configured to construct a circular region with the location information of the central key point as the center and the radius within a second preset distance range; and to obtain multiple candidate key points on the circular region corresponding to the central key point using a preset point selection method.

[0131] In some embodiments, the extension module is further configured to determine the screening results of multiple candidate key points for each central key point according to the sampling time sequence of the robot at multiple central key points and based on the position information of each central key point; and to determine the candidate key points corresponding to the screening results that meet the screening conditions as target key points; wherein the screening conditions include: there is no central key point within a third preset distance range of the candidate key points, or there are no candidate key points that have been determined as target key points and the central key point; and the candidate key points are not within the outline of obstacles in the constructed map.

[0132] In some embodiments, the extension module is further configured to take the target key point among multiple key points as the center key point, and repeatedly perform the steps of determining and filtering candidate key points until the number of iterations is within a preset range, thereby obtaining multiple target key points after iteration; and determine the multiple target key points after iteration and the multiple center key points as multiple key points, thereby obtaining the location information of multiple key points.

[0133] The descriptions of the apparatus embodiments above are similar to those of the method embodiments above, and have similar beneficial effects. In some embodiments, the functions or modules included in the apparatus provided in this application can be used to perform the methods described in the method embodiments above. For technical details not disclosed in the apparatus embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0134] It should be noted that, in the embodiments of this application, if the above-described data processing method is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of this application, or the part that contributes to the related technology, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, mobile hard drives, read-only memory (ROM), magnetic disks, or optical disks. Thus, the embodiments of this application are not limited to any specific hardware, software, or firmware, or any combination of hardware, software, and firmware.

[0135] This application provides a computer device including a memory and a processor. The memory stores a computer program that can run on the processor. When the processor executes the program, it implements some or all of the steps in the above-described method.

[0136] This application provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements some or all of the steps in the above-described method. The computer-readable storage medium can be transient or non-transient.

[0137] This application provides a computer program including computer-readable code. When the computer-readable code is run in a computer device, the processor in the computer device performs some or all of the steps for implementing the above-described method.

[0138] This application provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program. When the computer program is read and executed by a computer, it implements some or all of the steps in the above-described method. This computer program product can be implemented specifically through hardware, software, or a combination thereof. In some embodiments, the computer program product is specifically embodied as a computer storage medium; in other embodiments, the computer program product is specifically embodied as a software product, such as a software development kit (SDK), etc.

[0139] It should be noted that the descriptions of the various embodiments above tend to emphasize the differences between them, while their similarities or commonalities can be referred to interchangeably. The descriptions of the above embodiments of the device, storage medium, computer program, and computer program product are similar to the descriptions of the above method embodiments and have similar beneficial effects. For technical details not disclosed in the embodiments of the device, storage medium, computer program, and computer program product of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0140] Figure 12 This is a schematic diagram of the composition structure of a computer device provided in an embodiment of this application, such as... Figure 12 As shown, the computer device 1200 includes a memory 1210 and a processor 1220; wherein, the memory 1210 stores a computer program that can run on the processor 1220; when the processor 1220 executes the computer program, it implements the method provided in the above embodiments.

[0141] The memory 1210 stores computer programs that can run on the processor. The memory 1210 is configured to store instructions and applications that can be executed by the processor 1220. It can also cache data to be processed or already processed (e.g., image data, audio data, voice communication data and video communication data) in the processor 1220 and the various modules in the control device 1200. It can be implemented by flash memory or random access memory (RAM).

[0142] The processor 1220 implements the steps of any of the above control methods when executing a program. The processor 1220 typically controls the overall operation of the computer device 1200.

[0143] This application provides a computer storage medium that stores one or more programs that can be executed by one or more processors to implement the steps of the path planning method as described in any of the above embodiments.

[0144] It should be noted that the descriptions of the storage medium and device embodiments above are similar to the descriptions of the method embodiments above, and have similar beneficial effects. For technical details not disclosed in the storage medium and device embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0145] The aforementioned processor can be at least one of the following: Application Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), Central Processing Unit (CPU), Controller, Microcontroller, and Microprocessor. It is understood that other electronic devices can also implement the functions of the aforementioned processor, and this application does not specifically limit the specific implementation.

[0146] The aforementioned computer storage media / memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), magnetic random access memory (FRAM), flash memory, magnetic surface memory, optical disc, or compact disc read-only memory (CD-ROM), etc.; or it can be various terminals that include one or any combination of the above-mentioned memories, such as mobile phones, computers, tablet devices, personal digital assistants, etc.

[0147] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above steps / processes do not imply a sequential order of execution; the execution order of each step / process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the above embodiments of this application are merely descriptive and do not represent the superiority or inferiority of the embodiments.

[0148] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0149] The above description is merely an embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A path planning method, characterized in that, The path planning method includes: Obtain the robot's position information at multiple data sampling points during map construction; Based on the location information corresponding to the multiple data sampling points, multiple central key points are selected from the multiple data sampling points; the distance between any two central key points that are adjacent in sampling time is within a first preset distance range. Among the multiple key points obtained by expanding the central key point, the neighbor points of each key point are determined; the robot trajectory between the neighbor points and the key points does not interfere with the outline of obstacles in the constructed map; Based on the location information of the multiple key points and the location information of the neighboring points of each key point, the movement path of the robot from the current position to the target position is determined.

2. The path planning method according to claim 1, characterized in that, The step of determining the robot's movement path from its current position to its target position based on the location information of the multiple key points and the location information of the neighboring points of each key point includes: For each of the plurality of key points, a sub-path between the key point and each of the neighboring points is determined based on the location information of the key point and the location information of the neighboring points corresponding to the key point. Based on the multiple sub-paths, a movement path is determined for the robot to move from its current position to its target position.

3. The path planning method according to claim 2, characterized in that, The step of determining the movement path of the robot from its current position to its target position based on multiple sub-paths includes: The key point closest to the robot's current position is determined as the starting key point, and the key point closest to the robot's target position is determined as the ending key point. Based on the location information of the starting key point and the location information of the ending key point, a target sub-path is selected from multiple sub-paths; The robot's movement path is constructed based on the starting path between the robot's current position and the starting key point, the ending path between the robot's target position and the ending key point, and multiple target sub-paths.

4. The path planning method according to any one of claims 1 to 3, characterized in that, The method further includes: For each central key point, based on the location information of the central key point, multiple candidate key points whose distance from the central key point is within a second preset distance range are determined; Based on the location information of each central key point, multiple candidate key points corresponding to each central key point are filtered to obtain multiple target key points; The multiple target key points and the multiple center key points are identified as the multiple key points, and the location information of the multiple key points is obtained.

5. The path planning method according to claim 4, characterized in that, The step of determining multiple candidate key points within a second preset distance range from the central key point based on the location information of the central key point includes: Construct a circular region with the location information of the central key point as the center and a radius within the second preset distance range; Multiple candidate key points are obtained in the circular area corresponding to the central key point using a preset point selection method.

6. The path planning method according to claim 4, characterized in that, Based on the location information of each of the central key points, multiple candidate key points corresponding to each central key point are filtered to obtain multiple target key points, including: According to the sampling time sequence of the robot at multiple central key points, the screening results of multiple candidate key points for each central key point are determined sequentially based on the position information of each central key point; The candidate key points corresponding to the screening results that meet the screening criteria are determined as the target key points; The filtering criteria include: Within the third preset distance range of the candidate key points, there is no central key point, or there are no candidate key points and the central key point that have been identified as target key points; The candidate keypoints are not within the outline of obstacles in the constructed map.

7. The path planning method according to claim 4, characterized in that, The step of determining the plurality of target key points and the plurality of center key points as the plurality of key points and obtaining the location information of the plurality of key points includes: The target key point among the multiple key points is taken as the central key point, and the candidate key point determination and screening steps are repeated until the number of iterations is within the preset number of iterations, resulting in multiple target key points after multiple iterations. The target key points after the multiple cycles and the multiple center key points are determined as the multiple key points, and the position information of the multiple key points is obtained.

8. A path planning device, characterized in that, The path planning device includes: The acquisition module is used to acquire the location information of multiple data sampling points when the robot is building the map; The filtering module is used to filter multiple central key points from the multiple data sampling points based on the location information corresponding to each of the multiple data sampling points; the distance between any two central key points that are adjacent in sampling time among the multiple data sampling points is within a first preset distance range; The first determining module is used to determine the neighbor points of each key point among multiple key points obtained by expanding the central key point; the robot trajectory between the neighbor points and the key points does not interfere with the outline of obstacles in the constructed map; The second determining module is used to determine the movement path of the robot from the current position to the target position based on the position information of the plurality of key points and the position information of the neighboring points of each key point.

9. A computer device, characterized in that, include: Memory, used to store executable instructions; A processor, when executing executable instructions stored in the memory, implements the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The storage medium stores executable instructions that, when executed by a processor, implement the method described in any one of claims 1 to 7.

Images