Autonomous mapping method and device for mobile robot and electronic equipment

Abstract

The invention relates to an autonomous mapping method and device for a mobile robot and electronic equipment, and the method comprises the steps: building a passable path network of a target scene through scanning an architectural plane graph of the target scene; identifying all path closed loops in the passable path network; determining at least one group of path closed loop combinations meeting conditions in all the path closed loops; determining a target path closed-loop combination in the at least one group of path closed-loop combinations, and generating a guide path according to the target path closed-loop combination; and mapping the target scene according to radar scanning data when the mobile robot moves in the guide path to obtain an SLAM map. The mobile robot autonomous mapping method provided by the invention is high in efficiency and high in accuracy, and solves the problem of low mapping efficiency due to the fact that a current SLAM mapping mode mainly depends on manual operation of the mobile robot.

Description

Technical Field

[0001] The present application relates to the field of SLAM mapping, and in particular to a method, device and electronic equipment for autonomous mapping of a mobile robot. Background Art

[0002] Building an accurate and complete environmental map is essential for mobile robots with automatic navigation. Currently, manually operating mobile robots for map recording and mapping is a common method. However, large-scale industrial or commercial environments have high requirements for the integrity and accuracy of SLAM (Simultaneous Localization and Mapping) mapping, including full coverage of the map and high-precision positioning information of AGV. However, the current mainstream SLAM mapping method mainly relies on manual operation of mobile robots, which relies heavily on the operator's professional level of SLAM mapping. For general operators, it is quite difficult to achieve map information recording that meets the requirements of SLAM mapping. Since operators cannot evaluate the mapping effect and quality in real time, some map locations may be omitted or the amount of information is insufficient, resulting in incomplete maps; or, due to the failure to close the path in time during the recording of map information, the mapping error is large. When these problems occur, the mobile robot needs to be re-operated to walk to record the map information, which greatly reduces the efficiency of mapping.

[0003] The current SLAM mapping method mainly relies on manually controlled mobile robots, which results in low mapping efficiency. No effective solution has been proposed so far. Summary of the invention

[0004] The present invention provides a mobile robot autonomous mapping method, device and electronic equipment to solve the problem that the current SLAM mapping method mainly relies on manual control of mobile robots and has low mapping efficiency.

[0005] In a first aspect, the present invention provides a method for autonomous mapping of a mobile robot, wherein the mobile robot carries a radar, and the method for autonomous mapping comprises:

[0006] Establishing a navigable path network of the target scene by scanning a building plan of the target scene;

[0007] Identifying all path closures in the traversable path network;

[0008] At least one group of path closed loop combinations that meet conditions is determined among all the path closed loops, and the conditions that the path closed loop combinations need to meet include:

[0009] Condition 1: There is no overlapping inner loop area between any two path closed loops in the path closed loop combination;

[0010] Condition 2: the inner area of ​​any path closed loop in the path closed loop combination is covered by the scanning range of the radar running on the path closed loop;

[0011] Condition 3: the inner areas of all the path closed loops in the path closed loop combination cover the inner areas of all the path closed loops in the traversable path network;

[0012] Determine a target path closed-loop combination in at least one group of the path closed-loop combinations, and generate a guidance path according to the target path closed-loop combination;

[0013] The target scene is mapped according to the radar scanning data when the mobile robot moves in the guidance path to obtain a SLAM map.

[0014] In a second aspect, the present invention provides a mobile robot autonomous mapping device, wherein the mobile robot carries a radar, and the autonomous mapping device comprises:

[0015] A map scanning module, used to establish a navigable path network of the target scene by scanning the architectural plan of the target scene;

[0016] A path identification module, used to identify all path closed loops in the traversable path network;

[0017] The path determination module is used to determine at least one group of path closed loop combinations that meet conditions among all the path closed loops, and the conditions that the path closed loop combinations need to meet include:

[0018] Condition 1: There is no overlapping inner loop area between any two path closed loops in the path closed loop combination;

[0019] Condition 2: the inner area of ​​any path closed loop in the path closed loop combination is covered by the scanning range of the radar running on the path closed loop;

[0020] Condition 3: the inner areas of all the path closed loops in the path closed loop combination cover the inner areas of all the path closed loops in the traversable path network;

[0021] A path generation module, used to determine a target path closed-loop combination in at least one group of the path closed-loop combinations, and generate a guidance path according to the target path closed-loop combination;

[0022] The map building module is used to build a map of the target scene according to the radar scanning data when the mobile robot moves in the guidance path to obtain a SLAM map.

[0023] In a second aspect, the present invention provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the mobile robot autonomous mapping method described in the first aspect.

[0024] Compared with the related art, the present invention establishes a traversable path network of the target scene by scanning the architectural plan of the target scene, and determines the guidance path of the mobile robot based on the traversable path network, thereby realizing the autonomous movement of the robot. Moreover, radar scanning along the guidance path can cover the entire target scene. In addition, the guidance path is generated based on a path closure loop, and the path closure loop can eliminate the movement error of the mobile robot. Therefore, the mobile robot autonomous mapping method provided by the present invention is not only efficient but also accurate, and solves the problem that the current SLAM mapping method mainly relies on manual control of mobile robots and has low mapping efficiency.

[0025] Details of one or more embodiments of the present application are set forth in the following drawings and description to make other features, objects, and advantages of the present application more readily apparent. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 is a flow chart of a mobile robot autonomous mapping method provided in an embodiment of the present invention;

[0027] Figure 2 is a schematic diagram of the division of a building plan provided in an embodiment of the present invention;

[0028] Figure 3 is a schematic diagram of a navigable path network provided in an embodiment of the present invention;

[0029] Figure 4 is a schematic diagram of processing of a traversable path network endpoint provided in an embodiment of the present invention;

[0030] Figure 5 is a schematic diagram of public line merging provided in an embodiment of the present invention;

[0031] Figure 6 is a schematic diagram of a guiding path provided in an embodiment of the present invention;

[0032] Figure 7 It is a flow chart of a mobile robot autonomous mapping method provided in one embodiment of the present invention. DETAILED DESCRIPTION

[0033] In order to more clearly understand the purpose, technical solutions and advantages of the present application, the present application is described and illustrated below in conjunction with the accompanying drawings and embodiments.

[0034] Unless otherwise defined, the technical terms or scientific terms involved in this application shall have the general meaning understood by people with ordinary skills in the technical field to which this application belongs. The words "one", "a", "a", "the", "these" and the like in this application do not represent quantitative restrictions, and they can be singular or plural. The terms "include", "comprise", "have" and any variants thereof involved in this application are intended to cover non-exclusive inclusions; for example, a process, method and system, product or device comprising a series of steps or modules (units) is not limited to the listed steps or modules (units), but may include unlisted steps or modules (units), or may include other steps or modules (units) inherent to these processes, methods, products or devices. The words "connect", "connected", "coupled" and the like involved in this application are not limited to physical or mechanical connections, but may include electrical connections, whether directly or indirectly. The "multiple" involved in this application refers to two or more. "And / or" describes the association relationship of associated objects, indicating that there can be three relationships. For example, "A and / or B" can mean: A exists alone, A and B exist at the same time, and B exists alone. Usually, the character " / " indicates that the objects associated with each other are in an "or" relationship. The terms "first", "second", "third", etc. involved in this application are only used to distinguish similar objects and do not represent a specific ordering of the objects.

[0035] In an embodiment of the present invention, a method for autonomous mapping of a mobile robot is provided, wherein the mobile robot carries a radar. The method generates a moving path of the mobile robot so that the mobile robot can move autonomously. During the autonomous movement, the mobile robot collects environmental information through the radar, and finally uses the environmental information to perform SLAM (Simultaneous Localization and Mapping) mapping.

[0036] Figure 1 is a flow chart of a mobile robot autonomous mapping method provided in an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0037] Step S110 , establishing a navigable path network of the target scene by scanning the building plan of the target scene.

[0038] Among them, the target scene is the scene that needs to be mapped.

[0039] Specifically, the CAD map (building plan of the target scene) is divided into multiple rectangular areas, which represent the areas where unmanned vehicles (mobile robots) can pass. The overlap between these rectangular boxes represents accessibility. By analyzing the connectivity between each rectangular box, a topological relationship network between passable areas can be established to express the adjacent connectivity between them. Generate symmetry axes for each rectangular box, which can help determine the center and direction of the area. Through the topological structure between the rectangular boxes, the symmetry axes are extended so that the center lines with direct connectivity have intersections, thereby obtaining the center line of the entire pass road. These center lines can serve as the basis for path planning to help unmanned vehicles navigate and move in a warehousing environment. Figure 2 shown.

[0040] In order to determine the key points in path planning, the intersection points and endpoints are extracted from the center line. These key points can be used as reference points for path planning to help determine the turning and connection relationships of the path. Finally, the connectivity and distance between these key points are integrated into a topological relationship network, which is the traversable path network, such as Figure 3 This network can help analyze feasible paths and optimal paths in path planning and provide support for the navigation of unmanned vehicles.

[0041] In order to ensure that the target scene can be fully mapped, a multi-path generation strategy is designed. When generating the guidance path, multiple path strategies are considered, including full area coverage and closed loop priority. When the mobile robot travels a certain distance, it will be guided back to the specified starting position and then move towards the key points of the unknown area again. Steps S120 to S140 are the steps for generating the guidance path.

[0042] Step S120, identifying all path loops in the traversable path network.

[0043] Among them, the path closed loop includes at least three traversable path network intersections, for example, a path closed loop is: intersection A - intersection B - intersection C - intersection A.

[0044] In some embodiments, identifying all path closures in the traversable path network includes:

[0045] For any intersection in the traversable network, the intersection is used as the search starting point, and the path loop containing the intersection is determined by the breadth-first search algorithm. The specific process is: take a certain intersection as the search starting point, first determine the intersections adjacent to the search starting point, define them as first-level intersections, then determine the adjacent intersections of each first-level intersection, define them as second-level intersections, and so on, determine third-level intersections, fourth-level intersections, and so on. Thus, an intersection tree expanded with the search starting point can be obtained. In the intersection tree, the lower-level intersections are the adjacent intersections of the upper-level intersections to which they belong. Therefore, when the search starting point appears repeatedly in the intersection tree, the search starting point can be connected in series with the initial search starting point to obtain a path loop.

[0046] In summary, by determining the intersection tree of each intersection through the breadth-first search algorithm, all path closures containing the intersection can be obtained, and finally all path closures in the traversable path network can be determined.

[0047] Step S130: Determine at least one path closure combination that meets the conditions among all path closures. The conditions that the path closure combination needs to meet include:

[0048] Condition 1: There is no overlapping intra-loop area between any two path closed loops in the path closed loop combination.

[0049] Condition 2: The inner area of ​​any path closed loop in the path closed loop combination is covered by the scanning range of the radar running on the path closed loop.

[0050] Condition 3: The inner areas of all path closures in the path closure combination cover the inner areas of all path closures in the traversable path network.

[0051] The inner region of a closed path loop refers to the region enclosed by the closed loop. If there is no overlapping inner region between two closed path loops, it means that there is a gap between the two closed path loops or only some edges overlap.

[0052] The determination steps of condition two include: for any path loop in the path loop combination, the path loop is expanded according to the radar range. If the expanded area of ​​the path loop covers the inner area of ​​the path loop, condition two is determined to be met.

[0053] The expansion area of ​​a path closed loop refers to the scanning range of the radar when the mobile robot moves on the path closed loop. If the expansion area of ​​a path closed loop covers the area inside the loop, it means that the middle of the area inside the loop will also be scanned by the radar.

[0054] For step S130, all path closure combinations may be traversed first, and then each path closure combination may be judged in turn whether it satisfies the above three conditions, and the path closure combinations satisfying the conditions are retained. Accordingly, the path closures may be screened in the path closure set according to the above three conditions to form path closure combinations satisfying the conditions.

[0055] Therefore, in some embodiments, at least one set of path closure combinations that meet the conditions is determined among all path closures, including:

[0056] a. Initial closed-loop screening.

[0057] a). Traverse the path closed loop set and select a path closed loop as the starting closed loop of the current combination.

[0058] b) Determine whether the closed loop of the path satisfies complete coverage:

[0059] i. Expand the closed-loop boundary of the path according to the radar range to generate an expanded area.

[0060] ii. Determine whether the expansion area completely covers the inner area of ​​the path closed loop.

[0061] iii. If it is fully covered, then enter step b as the starting closed loop of the current combination; otherwise, continue to execute step a) in step a).

[0062] b. Add a new path closure to the current combination.

[0063] a). Select a candidate path loop from the remaining path loop set. If the remaining loop set is empty, proceed to step c.

[0064] b) Make the following judgments on the candidate path closure:

[0065] i. Determine whether the path loop meets the complete coverage requirement, which is the same as step b) in step a. If it is completely covered, proceed to the next step; if it is not completely covered, discard the path loop and repeat step a) in step b.

[0066] ii. Mutual exclusivity check: whether it is mutually exclusive with the path loops in the current combination. Through geometric algorithms (such as region intersection calculation), verify whether the candidate loop has no overlapping inner loop area with all path loops in the current combination.

[0067] c). If the above check is passed, the candidate path loop is added to the current combination; if the above check is not passed, the candidate path loop is not added; a recursive call is made.

[0068] c. Check the repeatability of path closed loop combination.

[0069] a). Sort the path closure combinations according to a certain rule (such as vertex order or index) to ensure that the same combination has a consistent representation.

[0070] b). If the closed-loop combination already exists, discard the closed-loop combination; otherwise, save the closed-loop combination.

[0071] Through the above steps, a series of path closed-loop combinations that meet the conditions and are not repeated can be finally obtained.

[0072] Step S140, determining a target path closed-loop combination in at least one group of path closed-loop combinations, and generating a guidance path according to the target path closed-loop combination.

[0073] If the path closures are grouped together, the group of path closures can be directly determined as the target path closure. If the path closures are multiple groups, one group of path closures needs to be selected from the multiple groups of path closures as the target path closure. The selection criteria are to facilitate the movement of the mobile robot, so a shorter path length and fewer turns are required.

[0074] Specifically, determining a target path closed-loop combination in at least one group of path closed-loop combinations includes:

[0075] The index value of each path closed-loop combination is calculated, and the path closed-loop combination with the smallest index value is determined as the target path closed-loop combination.

[0076] The calculation formula of the indicator value is:

[0077] Cost=Length_looppath*K_length+Number_turns*K_turns

[0078] Among them, Cost represents the index value, Length_looppath represents the total length of all path loops in the combination, Number_turns represents the total number of turns of all path loops in the combination, and K_length and K_turns both represent weights.

[0079] It should be noted that the nodes in the traversable path network include endpoints in addition to intersections. The endpoints in the target scene are the starting point or end point of the traversable path. The path loops in the target path loop combination above are all composed of intersections. If the mobile robot only moves between intersections, it may not be possible for the radar to scan the entire target scene. Therefore, a better choice is to connect each endpoint with the path loop closest to it, so that when the mobile robot moves on the path loop, it can go to the endpoint and return to the path loop.

[0080] Furthermore, in some embodiments, the autonomous mapping method further includes:

[0081] Reference Figure 4 , for any endpoint in the traversable network, determine the shortest path from the endpoint to the nearest path closure in the target path closure combination and add the shortest path to the nearest path closure.

[0082] The guidance path is the moving path of the mobile robot during scanning. Therefore, the starting point of the guidance path is the starting point of the map. Specifically, the guidance path is determined according to the closed-loop combination of the target path, including:

[0083] Taking the starting point of the map construction as the starting point of the guidance path, determine at least one candidate path for the mobile robot to sequentially walk through all the path closed loops in the target path closed loop combination from the starting point of the guidance path; and determine the guidance path from the candidate paths.

[0084] Exemplarily, the mobile robot first needs to go to the nearest path closed loop A in the target path closed loop combination, and then the mobile robot moves along the path closed loop A for one circle, and then goes to the adjacent path closed loop B and moves along the path closed loop B for one circle, and so on, the mobile robot completes all the path closed loops. The path that the mobile robot has walked is the guide path. This step simulates all feasible guidance paths for the mobile robot. It should be noted that if a path closed loop has an endpoint connected to it, the mobile robot needs to go to the endpoint when moving along the path closed loop. Exemplarily, there is an intersection a in the path closed loop A, and the endpoint b is connected to the intersection a (intersection a is the nearest closed loop point of endpoint b). When the mobile robot moves to the intersection a, it needs to go to the endpoint b and return to the intersection a and then continue to move on the path closed loop A.

[0085] There may be multiple paths that the robot can move. In this step, all the paths that the mobile robot can move are determined as candidate paths, and the candidate path that is most suitable for the mobile robot to move is selected as the guide path. The selection criteria can be the same as the selection criteria for the target path closed-loop combination. The index value of each candidate path can be calculated first, and the candidate path with the smallest index value can be used as the guide path. The index value calculation formula is as follows:

[0086] Cost=A*a+B*b

[0087] Among them, Cost represents the index value, A represents the length of the candidate path, B represents the total number of turns in the candidate path, and a and b both represent weights.

[0088] Reference Figure 5It should be noted that in the process of determining the guidance path, some linear paths (endpoint to closed loop point, or end point to map starting point) may have overlapping parts, and the overlapping parts can be merged. For example, there is point A reaching point C through point B, and there is point D reaching point C through point B, then there is an overlapping part from point B to point C. Assuming point C is a closed loop point, point A and point D are end points, when the mobile robot moves to point C, it needs to go to point A and point D and return. At this time, the mobile robot can first go to point A and then return to point B (not point C), and then go to point D and return to point C through point B.

[0089] Reference Figure 6 , which is a generated guidance path in one embodiment, and the numbers in the figure represent the movement sequence of the mobile robot.

[0090] Step S150, mapping the target scene according to the radar scanning data when the mobile robot moves in the guidance path to obtain a SLAM map.

[0091] Specifically, online SLAM is a complex and critical technology that provides important support for autonomous mobile robots to achieve positioning and map construction in unknown environments. First, mobile robots are equipped with various sensors, such as lidar, cameras, etc., which perceive the surrounding environment and their own status in real time. The SLAM system uses sensor data for feature extraction and matching. It analyzes sensor data, identifies feature points or feature areas in the environment, and matches them with known features in the map to determine the current position of the AGV and the update information of the map.

[0092] Based on feature matching, the SLAM system builds an environmental map in real time and continuously updates the map information based on sensor data. This process requires efficient mapping algorithms and data fusion technology to ensure the accuracy and consistency of the map. At the same time, the SLAM system estimates and optimizes the robot's state through probabilistic filters and graph optimization methods.

[0093] During the operation of the SLAM system, closed loops are constantly detected, and the map and robot status are corrected through closed loop search to reduce error accumulation and improve map consistency. This process is crucial for long-term SLAM systems and can effectively improve the stability and reliability of map construction. Every time the mobile robot travels a certain distance or turns more than a certain angle, a new submap is created and loops are detected in the historical submap to timely correct the submap's posture and update the global map.

[0094] The currently used mapping algorithm is developed based on the core idea of ​​separation of front-end and back-end. It is mainly divided into modules such as sensor data input, data preprocessing, pose estimation, front-end matching, and back-end optimization. It provides real-time global mapping and positioning through rolling iteration of the optimization algorithm and deploying it in an embedded system environment.

[0095] In some of the embodiments, the autonomous mapping method further includes:

[0096] The intersections and / or endpoints in the traversable path network are used as key points, and the key points are used as alignment references to align the traversable path network with the SLAM map.

[0097] Specifically, before starting to generate guide lines, ensure that the coordinate systems of the building plan and the SLAM map are accurately aligned. The alignment process can consider the matching of global feature points, such as fixed walls, corners, or other significant landmarks. Select 3 or more corresponding points on the building plan map and the SLAM submap respectively, and then use the matching algorithm to generate the transformation matrix between the two.

[0098] In addition, the traversable path network is divided into multiple sub-areas, and reference markers are set at key points for registration with the real-time map generated by SLAM to ensure consistency between area coverage and guidance paths.

[0099] In some of the embodiments, the autonomous mapping method further includes:

[0100] In real time, the SLAM map is used to determine whether there are key points in the target path where the environment has been constructed but the mobile robot has not been to; if so, the target path is updated based on the key points where the environment has not been constructed and the mobile robot has not been to.

[0101] During the mapping process, the guidance route is dynamically adjusted based on the actual mapping effect. For example, when the number of valid point clouds around a key point reaches a certain standard, it can be determined that the area already has the necessary environmental information and the robot does not need to go there. This dynamic adjustment mechanism can improve efficiency, avoid unnecessary repeated paths, and ensure the rational use of resources.

[0102] In summary, the present invention establishes a traversable path network of the target scene by scanning the architectural plan of the target scene, and determines the guidance path of the mobile robot based on the traversable path network, thereby realizing the autonomous movement of the robot. Moreover, radar scanning along the guidance path can cover the entire target scene. In addition, the guidance path is generated based on a path closure loop, and the path closure loop can eliminate the movement error of the mobile robot. Therefore, the mobile robot autonomous mapping method provided by the present invention is not only efficient but also accurate, and solves the problem that the current SLAM mapping method mainly relies on manual control of mobile robots and has low mapping efficiency.

[0103] The mobile robot autonomous mapping method provided by the present invention is described below through a specific embodiment. The mobile robot autonomous mapping method is executed by a computer processor, which is a software system.

[0104] In a specific embodiment, the mobile robot autonomous mapping method includes: a map preprocessing module, a relocalization module, an online mapping module, a map alignment module, a guidance path generation module, a motion control module, and a visualization and interaction module.

[0105] Reference Figure 7 , which is a flow chart of the autonomous mapping method of the mobile robot in this embodiment.

[0106] 1. Map preprocessing module.

[0107] In the CAD map (building floor plan), it is divided into multiple rectangular areas, which represent the areas where unmanned vehicles can pass. The overlap between these rectangular boxes represents accessibility. By analyzing the connectivity relationship between each rectangular box, a topological relationship network between passable areas can be established to express the adjacent connectivity relationship between them. Generate symmetry axes for each rectangular box, which can help determine the center and direction of the area. Through the topological structure between the rectangular boxes, the symmetry axes are extended so that the center lines with direct connectivity have intersections, thereby obtaining the center line of the entire pass road. These center lines can serve as the basis for path planning to help unmanned vehicles navigate and move in a warehousing environment. Figure 2 shown.

[0108] In order to determine the key points in path planning, the intersections and endpoints are extracted from the centerline. These key points can be used as reference points for path planning to help determine the turning and connection relationships of the path. Finally, the connectivity and distance between these key points are integrated into a topological relationship network, such as Figure 3 This network can help analyze feasible paths and optimal paths in path planning and provide support for the navigation of unmanned vehicles.

[0109] 2. Map alignment module.

[0110] Precise alignment: Before starting to generate guide lines, ensure that the coordinate systems of the offline building plan map and the SLAM map are accurately aligned. The alignment process can consider the matching of global feature points, such as fixed walls, corners, or other significant landmarks. Select 3 or more corresponding points on the construction plan map and the SLAM submap respectively, and then use the matching algorithm to generate the transformation matrix between the two.

[0111] Region division and reference point calibration: Divide the CAD map into multiple sub-regions and set reference markers at key points for registration with the real-time map generated by SLAM to ensure consistency between region coverage and guide lines.

[0112] 3. Guidance path generation module.

[0113] In the guidance path generation stage under the global environment, the offline building floor plan and marked passable areas are used, the positioning information provided by the positioning radar and the real-time generated SLAM map are used to dynamically plan a series of guidance paths connecting key points to guide the mobile robot to drive to areas that have not yet been covered by the map.

[0114] In addition, to ensure that the system can fully map the environment, a multiple path generation strategy is designed. When generating guide lines, multiple path strategies are considered, including full area coverage and closed loop priority. When the mobile robot travels a certain distance, the system will guide it back to the specified starting position and then move towards the key points of the unknown area again. First, we plan a basic closed-loop path based on the starting point of the mobile robot, which satisfies certain constraints such as distance and number of turns. Every time the mobile robot drives to a key point along the guided closed-loop path, the system will update the closed-loop path. When updating the closed-loop path, the system will be guided by the unexplored area and consider constraints such as distance, number of turns, and coverage area. When the map is explored, the mobile robot will return to the original starting point to meet the large loop condition of the overall path.

[0115] 3. Online mapping module.

[0116] Online SLAM is a complex and critical technology that provides important support for autonomous mobile robots to achieve positioning and map building in unknown environments. First, mobile robots are equipped with various sensors, such as lidar, cameras, etc., which perceive the surrounding environment and their own status in real time. The SLAM system uses sensor data for feature extraction and matching. It analyzes sensor data, identifies feature points or feature areas in the environment, and matches them with known features in the map to determine the current position of the AGV and update information of the map.

[0117] Based on feature matching, the SLAM system builds an environmental map in real time and continuously updates the map information based on sensor data. This process requires efficient mapping algorithms and data fusion technology to ensure the accuracy and consistency of the map. At the same time, the SLAM system estimates and optimizes the robot's state through probabilistic filters and graph optimization methods.

[0118] During the operation of the SLAM system, closed loops are constantly detected, and the map and robot status are corrected through closed loop search to reduce error accumulation and improve map consistency. This process is crucial for long-term SLAM systems and can effectively improve the stability and reliability of map construction. Every time the mobile robot travels a certain distance or turns more than a certain angle, a new submap is created and loops are detected in the historical submap to timely correct the submap's posture and update the global map.

[0119] The currently used mapping algorithm is developed based on the core idea of ​​separation of front-end and back-end. It is mainly divided into modules such as sensor data input, data preprocessing, pose estimation, front-end matching, and back-end optimization. It provides real-time global mapping and positioning through rolling iteration of the optimization algorithm and deploying it in an embedded system environment.

[0120] 4. Visualization module.

[0121] Grid coverage detection and display: Divide the map into grids and monitor the coverage of each grid in real time. If some grid areas are not covered by SLAM mapping, the system will prompt to redirect the vehicle or supplement the mapping.

[0122] Area coverage heat map: Use heat map to display the covered and uncovered areas in the SLAM map in real time to ensure that the guide line leads the vehicle to cover all areas. If unreasonable places are found, users can immediately manually adjust the vehicle's movement path locally and confirm the correction effect.

[0123] Phased confirmation and verification: Users can confirm the integrity and accuracy of the generated map area by area, and then move on to the next area after confirming that the requirements are met, ensuring the integrity and high quality of the final map.

[0124] Local incremental mapping: For uncovered areas or areas with large errors, the local re-mapping function allows users to refine or expand the map generated by SLAM through manual operations.

[0125] In an embodiment of the present invention, a mobile robot autonomous mapping device is also provided, and the mobile robot carries a radar. The device is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. The terms "module", "unit", "subunit", etc. used below can implement a combination of software and / or hardware for a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0126] Autonomous mapping device includes:

[0127] A map scanning module is used to establish a navigable path network of the target scene by scanning the building plan of the target scene;

[0128] A path identification module is used to identify all path closures in the traversable path network;

[0129] The path determination module is used to determine at least one path closed loop combination that meets the conditions among all path closed loops. The conditions that the path closed loop combination needs to meet include:

[0130] Condition 1: There is no overlapping inner loop area between any two path closed loops in the path closed loop combination;

[0131] Condition 2: The inner area of ​​any path closed loop in the path closed loop combination is covered by the scanning range of the radar running on the path closed loop;

[0132] Condition 3: The inner areas of all path closures in the path closure combination cover the inner areas of all path closures in the traversable path network;

[0133] A path generation module, used to determine a target path closed-loop combination from at least one group of path closed-loop combinations, and generate a guidance path according to the target path closed-loop combination;

[0134] The map building module is used to build a SLAM map for the target scene based on the radar scanning data when the mobile robot moves in the guidance path.

[0135] In an embodiment of the present invention, an electronic device is also provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to execute the mobile robot autonomous mapping method provided by the present invention.

[0136] It should be noted that the above modules can be functional modules or program modules, and can be implemented by software or hardware. For modules implemented by hardware, the above modules can be located in the same processor; or the above modules can be located in different processors in any combination.

[0137] It should be understood that the specific embodiments described herein are only used to explain the application, rather than to limit it. Based on the embodiments provided in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the protection scope of this application.

[0138] Obviously, the drawings are only some examples or embodiments of the present application. For ordinary technicians in the field, the present application can also be applied to other similar situations based on these drawings without creative work. In addition, it is understandable that although the work done in this development process may be complicated and lengthy, for ordinary technicians in the field, certain changes in design, manufacturing or production based on the technical content disclosed in this application are only conventional technical means and should not be regarded as insufficient content disclosed in this application.

Claims

1. A method for autonomous mapping by a mobile robot, wherein the mobile robot carries a radar, characterized in that: The autonomous mapping method comprises: Establishing a navigable path network of the target scene by scanning a building plan of the target scene; Identifying all path closures in the traversable path network; At least one group of path closed loop combinations that meet conditions is determined among all the path closed loops, and the conditions that the path closed loop combinations need to meet include: Condition 1: There is no overlapping inner loop area between any two path closed loops in the path closed loop combination; Condition 2: the inner area of ​​any path closed loop in the path closed loop combination is covered by the scanning range of the radar running on the path closed loop; Condition 3: the inner areas of all the path closed loops in the path closed loop combination cover the inner areas of all the path closed loops in the traversable path network; Determine a target path closed-loop combination in at least one group of the path closed-loop combinations, and generate a guidance path according to the target path closed-loop combination; The target scene is mapped according to the radar scanning data when the mobile robot moves in the guidance path to obtain a SLAM map.

2. The mobile robot autonomous mapping method according to claim 1, characterized in that: The identifying all closed path loops in the traversable path network includes: For any intersection in the traversable network, the intersection is used as a search starting point, and a path closed loop including the intersection is determined by a breadth-first search algorithm.

3. The mobile robot autonomous mapping method according to claim 1, characterized in that: The determination steps of the second condition include: For any path closed loop in the path closed loop combination, the path closed loop is expanded according to the radar range. If the expanded area of ​​the path closed loop covers the inner area of ​​the path closed loop, it is determined that the second condition is met.

4. The mobile robot autonomous mapping method according to claim 1, characterized in that: Determining a target path closed-loop combination from at least one group of the path closed-loop combinations includes: The index value of each group of the path closed loop combination is calculated, and the calculation formula of the index value is: Cost=Length_looppath*K_length+Number_turns*K_turns Among them, Cost represents the index value, Length_looppath represents the total length of all path loops in the combination, Number_turns represents the total number of turns of all path loops in the combination, and K_length and K_turns both represent weights; The path closed-loop combination with the smallest index value is determined as the target path closed-loop combination.

5. The mobile robot autonomous mapping method according to claim 2, characterized in that: The autonomous mapping method further comprises: For any endpoint in the traversable network, the shortest path from the endpoint to the nearest path closed loop in the target path closed loop combination is determined and the shortest path is added to the nearest path closed loop.

6. The mobile robot autonomous mapping method according to claim 2, characterized in that: Determining a guidance path according to the target path closed-loop combination includes: Taking the starting point of the map construction as the starting point of the guidance path, determining at least one candidate path for the mobile robot to sequentially walk through all the path closed loops in the target path closed loop combination starting from the starting point of the guidance path; The guidance path is determined from the candidate paths.

7. The mobile robot autonomous mapping method according to claim 2, characterized in that: The autonomous mapping method further comprises: The intersection points and / or endpoints in the traversable path network are used as key points, and the key points are used as alignment references to align the traversable path network with the SLAM map.

8. The mobile robot autonomous mapping method according to claim 7, characterized in that: The autonomous mapping method further comprises: Determine in real time according to the SLAM map whether there is a key point in the target path where the environment has been constructed and the mobile robot has not yet gone; If so, the target path is updated based on key points where the environment is not constructed and the mobile robot has not traveled.

9. A mobile robot autonomous mapping device, the mobile robot carrying a radar, characterized in that: The autonomous mapping device comprises: A map scanning module, used to establish a navigable path network of the target scene by scanning the architectural plan of the target scene; A path identification module, used to identify all closed path loops in the traversable path network; The path determination module is used to determine at least one group of path closed loop combinations that meet conditions among all the path closed loops, and the conditions that the path closed loop combinations need to meet include: Condition 1: There is no overlapping inner loop area between any two path closed loops in the path closed loop combination; Condition 2: the inner area of ​​any path closed loop in the path closed loop combination is covered by the scanning range of the radar running on the path closed loop; Condition 3: the inner areas of all the path closed loops in the path closed loop combination cover the inner areas of all the path closed loops in the traversable path network; A path generation module, used to determine a target path closed-loop combination in at least one group of the path closed-loop combinations, and generate a guidance path according to the target path closed-loop combination; The map building module is used to build a map of the target scene according to the radar scanning data when the mobile robot moves in the guidance path to obtain a SLAM map.

10. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to run the computer program to execute the mobile robot autonomous mapping method according to any one of claims 1 to 8.

Images