A mobile robot repositioning method, device and mobile robot

Abstract

The application relates to a mobile robot repositioning method and device and a mobile robot, wherein the method comprises the following steps: matching an environment image collected by a visual sensor in a prior map to complete visual repositioning and obtain a rough repositioning pose; when the visual repositioning fails, matching an environment point cloud collected by a laser sensor in the prior map to complete laser repositioning and obtain a rough repositioning pose; and extracting a local map on the prior map according to the rough repositioning pose, and performing point cloud registration on the current environment point cloud and the local map to obtain a precise repositioning pose. The application can guarantee repositioning accuracy and improve repositioning robustness.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a method, apparatus and mobile robot for relocation of a mobile robot. Background Technology

[0002] With the development of artificial intelligence technology, the research and development of mobile robots has been very rapid, and some service robots have gradually entered people's daily lives. Regardless of the type of robot, as long as it moves autonomously, it needs to navigate and locate itself in its working environment, becoming familiar with its surroundings and location.

[0003] In existing technologies, some robots are equipped with vision and laser sensors, combined with real-time localization and mapping technologies, which enable them to build environmental maps in unknown environments when their own position is uncertain, and use these environmental maps for autonomous localization and navigation.

[0004] In practical applications, robots may encounter situations such as being abducted or restarting. In these cases, the robot needs to relocalize. Currently, robots relocalize based on a pre-built environmental map, using visual or laser sensors. However, this single-sensor-based relocalization method has poor robustness and is easily affected by environmental changes such as lighting, weather, and seasons. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a mobile robot repositioning method, device and mobile robot to ensure repositioning accuracy and improve repositioning robustness.

[0006] The technical solution adopted by this invention to solve its technical problem is: to provide a mobile robot relocation method, comprising the following steps:

[0007] The environmental images acquired by the visual sensor are matched with the prior map to complete the visual relocalization and obtain the coarse pose of the relocalization.

[0008] When visual relocalization fails, the environmental point cloud collected by the laser sensor is matched with the prior map to complete laser relocalization and obtain the coarse pose of relocalization.

[0009] Based on the coarse repositioning pose, a local map is extracted from the prior map, and point cloud registration is performed between the current environment point cloud and the local map to obtain the fine repositioning pose.

[0010] The prior map includes a pre-constructed visual map and a laser map.

[0011] The environmental image acquired by the visual sensor is matched with a prior map to complete visual relocalization and obtain a coarse pose for relocalization. Specifically:

[0012] Extract the bag-of-words vector from the environmental image;

[0013] The bag-of-words vector is compared with each bag-of-words vector in the bag-of-words database in the visual map, the bag-of-words vector in the bag-of-words database that is closest to the bag-of-words vector is found, and the image corresponding to the closest bag-of-words vector is obtained.

[0014] The visual keyframes of the image are used as the coarse pose for relocalization.

[0015] The step of comparing the bag-of-words vector with each bag-of-words vector in the bag-of-words database of the visual map, and finding the bag-of-words vector in the database that is closest to the bag-of-words vector, specifically involves:

[0016] Calculate the similarity distance between the bag-of-words vectors and each bag-of-words vector in the bag-of-words database in the visual map;

[0017] Select the bag-of-words vector corresponding to the minimum value from the obtained similarity distances, and take this bag-of-words vector as the closest bag-of-words vector.

[0018] The environmental point cloud acquired by the laser sensor is matched with a prior map to complete laser relocalization and obtain a coarse pose for relocalization. Specifically:

[0019] Extract the scancontext descriptor from the environmental point cloud;

[0020] The scancontext descriptor is compared with each scancontext descriptor in the scancontext database of the laser map, and the scancontext descriptor in the scancontext database that is closest to the scancontext descriptor is found, and the laser point cloud corresponding to the closest scancontext descriptor is obtained.

[0021] The laser keyframes of the laser point cloud are used as the coarse pose for relocalization.

[0022] The step of comparing the scan context descriptor with each scan context descriptor in the scan context database of the laser map, and finding the scan context descriptor in the scan context database that is closest to the scan context descriptor, specifically involves:

[0023] Extract sub-descriptors from each scancontext descriptor in the scancontext database, and find candidate matching frames based on the sub-descriptors;

[0024] Calculate the similarity distance between the scancontext descriptor and the scancontext descriptors of each candidate matching frame;

[0025] Find the scancontext descriptor of the candidate matching frame that is closest to the scancontext descriptor. If the similarity distance between the two is less than a threshold, then the scancontext descriptor of the closest candidate matching frame is taken as the closest scan context descriptor.

[0026] The step of extracting a local map from the prior map based on the coarse repositioning pose, and performing point cloud registration between the current environment point cloud and the local map to obtain the fine repositioning pose specifically includes:

[0027] The laser point clouds corresponding to the laser point cloud keyframes whose coarse poses for distance relocalization are within a preset range are combined together and downsampled to obtain a local map.

[0028] A registration algorithm is used to register the currently observed environmental point cloud with the local map to obtain the fine pose for relocalization. When performing point cloud registration, the coarse pose for relocalization is used as the initial pose of the current environmental point cloud.

[0029] The technical solution adopted by this invention to solve its technical problem is: to provide a mobile robot repositioning device, comprising:

[0030] The visual relocalization module is used to match the environmental image acquired by the visual sensor with the prior map to complete the visual relocalization and obtain the coarse pose of the relocalization.

[0031] The laser repositioning module is used to match the environmental point cloud collected by the laser sensor with the prior map when the visual repositioning module fails to reposition, and to complete the laser repositioning and obtain the coarse pose of the repositioning.

[0032] The point cloud registration module is used to extract a local map on the prior map based on the coarse relocation pose, and perform point cloud registration between the current environment point cloud and the local map to obtain the fine relocation pose.

[0033] The prior map includes a pre-constructed visual map and a laser map.

[0034] The visual relocalization module includes:

[0035] A bag-of-words vector extraction unit is used to extract the bag-of-words vectors of the environmental image;

[0036] The bag-of-words vector comparison unit is used to compare the bag-of-words vector with each bag-of-words vector in the bag-of-words database in the visual map, find the bag-of-words vector in the bag-of-words database that is closest to the bag-of-words vector, and obtain the image corresponding to the closest bag-of-words vector.

[0037] The first determining unit is used to use the visual image keyframes of the image as the coarse pose for relocalization.

[0038] The laser repositioning module includes:

[0039] The descriptor extraction unit is used to extract the scancontext descriptor of the environmental point cloud;

[0040] The descriptor comparison unit is used to compare the scancontext descriptor with each scancontext descriptor in the scancontext database of the laser map, find the scancontext descriptor in the scancontext database that is closest to the scancontext descriptor, and obtain the laser point cloud corresponding to the closest scancontext descriptor.

[0041] The second determining unit is used to use the laser keyframe of the laser point cloud as the coarse pose for relocation.

[0042] The technical solution adopted by the present invention to solve its technical problem is: to provide a mobile robot, including a memory and a processor, wherein the memory is used to store computer execution instructions; the processor is used to execute the computer execution instructions stored in the memory, so that the mobile robot performs the above-mentioned mobile robot relocation method.

[0043] Beneficial effects

[0044] By employing the above-mentioned technical solutions, this invention has the following advantages and positive effects compared with existing technologies: Firstly, this invention rapidly matches environmental images acquired by a visual sensor with a visual map to obtain a coarse repositioning pose, or rapidly matches environmental point clouds acquired by a laser sensor with a laser map to obtain a coarse repositioning pose. Then, based on the coarse repositioning pose, a local map is extracted from the prior map, and the current environmental point cloud and the local map are registered using ICP or NDT algorithms to obtain a more accurate fine repositioning pose. Simultaneously, using methods based on both visual and laser sensors to obtain the coarse repositioning pose helps improve the robustness of the repositioning process. Attached Figure Description

[0045] Figure 1 This is a flowchart of the mobile robot relocation method according to the first embodiment of the present invention. Detailed Implementation

[0046] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, it should be understood that after reading the teachings of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent forms also fall within the scope defined by the appended claims.

[0047] The first embodiment of the present invention relates to a mobile robot relocation method, such as... Figure 1 As shown, it includes the following steps:

[0048] The visual relocalization step involves matching the environmental image acquired by the visual sensor with the prior map to complete the visual relocalization and obtain the coarse pose of the relocalization.

[0049] The laser relocalization step is initiated when the visual relocalization fails. It matches the environmental point cloud collected by the laser sensor with the prior map to complete the laser relocalization and obtain the coarse pose of the relocalization.

[0050] The point cloud registration step involves extracting a local map from the prior map based on the coarse relocation pose, and then registering the current environment point cloud with the local map to obtain the fine relocation pose.

[0051] The visual relocalization method is as follows: information is extracted from the environmental image currently acquired by the robot to obtain bag-of-words vectors. The bag-of-words vectors are compared with each bag-of-words vector in the bag-of-words database in the prior map. The bag-of-words vector in the database that is closest to the bag-of-words vector is found, and the visual image keyframe of the image corresponding to the closest bag-of-words vector is used as the coarse pose for relocalization.

[0052] The laser relocalization method is as follows: extract the scancontext descriptor from the laser point cloud currently acquired by the robot, compare the scancontext descriptor with each scancontext descriptor in the scancontext database in the prior map, find the scancontext descriptor in the scancontext database that is closest to the scancontext descriptor, and use the laser keyframe of the laser point cloud corresponding to the closest scancontext descriptor as the coarse pose for relocalization.

[0053] After obtaining the coarse pose of repositioning based on visual repositioning or laser repositioning, point cloud registration is performed, as follows: a local map of a certain size is extracted from the prior map, and the environmental point cloud collected by the robot is registered with the local map using the ICP or NDT algorithm to verify the accuracy of the coarse pose of repositioning and to obtain a more accurate fine pose of repositioning. This fine pose of repositioning is then used as the robot's final pose.

[0054] Among them, the prior map is a pre-built environment map, which includes a visual map and a laser map. The visual map consists of visual keyframes and a visual bag-of-words database, while the laser map consists of laser keyframes and a laser point cloud scancontext database.

[0055] The method provided in this embodiment can be applied to autonomous mobile robots, primarily for robot relocalization, i.e., re-determining the robot's initial position. In this embodiment, the robot can move autonomously, requiring navigation and positioning during movement. In practical applications, situations such as "kidnapping" or "rebooting" may occur. In these cases, the robot may lack or lose its previous position information, thus requiring re-determining the robot's current pose, i.e., relocalization. In this embodiment, the robot's laser sensor and vision sensor are combined, utilizing a prior map for robot relocalization. The vision sensor continuously acquires images of the robot's surrounding environment, and correspondingly, the laser sensor continuously acquires point clouds of the robot's surrounding environment.

[0056] When robot relocalization is required, the environmental image captured by the visual sensor at the robot's current position can be matched with the visual map. If the match fails, the environmental point cloud captured by the laser sensor at the robot's current position can be matched with the laser map. The pose of the visual keyframe matched with the current environmental image or the laser keyframe matched with the current environmental point cloud can be used as the coarse relocalization pose. Then, based on the coarse relocalization pose, a local map is extracted from the prior map, and point cloud registration is performed between the environmental point cloud captured at the current position and the local map, thereby obtaining a more accurate fine relocalization pose.

[0057] If both visual and laser relocalization fail, the robot can move to the next location and attempt relocalization again. The next location can be random or a location with rich features, increasing the success rate of the next relocalization attempt.

[0058] In this embodiment, visual relocalization employs the bag-of-words model technique. First, feature points are extracted from the current environment image. Then, the word corresponding to each feature point is found in the bag-of-words model dictionary, thus forming a bag-of-words vector. The bag-of-words vector of the current environment image is compared with each bag-of-words vector in the prior map's bag-of-words database. The similarity distance between them is calculated, and the bag-of-words vector in the database that is closest in distance to the bag-of-words vector of the current environment image is found. The keyframe of the visual image corresponding to this closest bag-of-words vector is then used as the coarse pose for relocalization.

[0059] In this embodiment, laser relocalization employs scancontext technology. Scancontext is a global descriptor used to describe 3D point clouds. It divides the point cloud into radial rings and circumferential sectors on a top-down view. Each sector is represented by the z-axis coordinate of the highest point within that region, ultimately forming a two-dimensional matrix-like global descriptor. This embodiment uses a two-step search algorithm for matching. First, sub-descriptors are extracted from each scancontext descriptor in the database, and candidate matching frames are found based on these sub-descriptors. Then, the similarity distance between the current environment point cloud's scancontext descriptor and the scancontext descriptors of each candidate matching frame is calculated. The scancontext descriptor of the candidate matching frame that is closest to the current environment point cloud's scancontext descriptor is found. If the similarity distance is less than a threshold, the keyframe of the laser point cloud corresponding to the scancontext descriptor of this closest candidate matching frame is used as the coarse pose for relocalization.

[0060] In this embodiment, point cloud registration can employ either the ICP or NDT algorithm. First, based on the coarse repositioning pose obtained from visual or laser repositioning, a local map is extracted from the prior map. Specifically, laser point clouds corresponding to keyframes within a preset range of the coarse repositioning pose are combined and downsampled to form a local map. Next, the environmental point cloud observed by the current laser sensor is registered with the local map. Registration can use either the ICP or NDT algorithm, where the coarse repositioning pose serves as the initial pose of the current environmental point cloud. If the ICP or NDT algorithm converges, the calculated result is the fine repositioning pose, which is the robot's final pose. If the ICP or NDT algorithm fails to converge within a certain number of iterations, the point cloud registration verification fails, i.e., repositioning fails, and the robot moves to the next location to reposition again.

[0061] It is easy to see that this invention first obtains a coarse pose for relocalization by quickly matching the environmental image acquired by a visual sensor with a visual map, or by quickly matching the environmental point cloud acquired by a laser sensor with a laser map. Then, based on the coarse pose, a local map is extracted from the prior map, and the current environmental point cloud and the local map are registered using ICP or NDT algorithms to obtain a more accurate fine pose for relocalization. Simultaneously using methods based on both visual and laser sensors to obtain the coarse pose for relocalization helps improve the robustness of the relocalization process.

[0062] A second embodiment of the present invention relates to a mobile robot repositioning device, comprising:

[0063] The visual relocalization module is used to match the environmental image acquired by the visual sensor with the prior map to complete the visual relocalization and obtain the coarse pose of the relocalization.

[0064] The laser repositioning module is used to match the environmental point cloud collected by the laser sensor with the prior map when the visual repositioning module fails to reposition, and to complete the laser repositioning and obtain the coarse pose of the repositioning.

[0065] The point cloud registration module is used to extract a local map on the prior map based on the coarse relocation pose, and perform point cloud registration between the current environment point cloud and the local map to obtain the fine relocation pose.

[0066] The prior map includes a pre-constructed visual map and a laser map.

[0067] The visual relocalization module includes:

[0068] A bag-of-words vector extraction unit is used to extract the bag-of-words vectors of the environmental image;

[0069] The bag-of-words vector comparison unit is used to compare the bag-of-words vector with each bag-of-words vector in the bag-of-words database in the visual map, find the bag-of-words vector in the bag-of-words database that is closest to the bag-of-words vector, and obtain the image corresponding to the closest bag-of-words vector.

[0070] The first determining unit is used to use the visual image keyframes of the image as the coarse pose for relocalization.

[0071] The bag-of-words vector comparison unit includes:

[0072] The first calculation subunit is used to calculate the similarity distance between the bag-of-words vector and each bag-of-words vector in the bag-of-words database in the visual map;

[0073] The selection unit is used to select the bag-of-words vector corresponding to the minimum value from the obtained similarity distances, and to take this bag-of-words vector as the closest bag-of-words vector.

[0074] The laser repositioning module includes:

[0075] The descriptor extraction unit is used to extract the scancontext descriptor of the environmental point cloud;

[0076] The descriptor comparison unit is used to compare the scancontext descriptor with each scancontext descriptor in the scancontext database of the laser map, find the scancontext descriptor in the scancontext database that is closest to the scancontext descriptor, and obtain the laser point cloud corresponding to the closest scancontext descriptor.

[0077] The second determining unit is used to use the laser keyframe of the laser point cloud as the coarse pose for relocation.

[0078] The descriptor comparison unit includes:

[0079] The sub-descriptor extraction sub-unit is used to extract sub-descriptors from each scancontext descriptor in the database and find candidate matching frames based on the sub-descriptors;

[0080] The second calculation subunit is used to calculate the similarity distance between the scancontext descriptor and the scancontext descriptors of each candidate matching frame;

[0081] A sub-unit is determined to find the scancontext descriptor of the candidate matching frame that is closest to the scancontext descriptor. If the similarity distance between the two is less than a threshold, the scancontext descriptor of the closest candidate matching frame is taken as the closest scancontext descriptor.

[0082] The point cloud registration module includes:

[0083] The local map generation unit is used to combine the laser point clouds corresponding to the laser point cloud keyframes with coarse poses of distance relocalization within a preset range, and perform downsampling operations to obtain a local map.

[0084] The point cloud registration unit uses a registration algorithm to register the currently observed environmental point cloud with the local map to obtain the fine pose for relocalization. When performing point cloud registration, the coarse pose for relocalization is used as the initial pose of the current environmental point cloud.

[0085] The third embodiment of the present invention relates to a mobile robot, including a memory and a processor, wherein the memory is used to store computer execution instructions; and the processor is used to execute the computer execution instructions stored in the memory, causing the mobile robot to perform the mobile robot relocation method of the first embodiment.

[0086] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0087] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0088] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0089] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the invention.

[0090] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A mobile robot relocalization method, characterized in that, Includes the following steps: The environmental images acquired by the visual sensor are matched with the prior map to complete the visual relocalization and obtain the coarse pose of the relocalization. When visual relocalization fails, the environmental point cloud collected by the laser sensor is matched with the prior map to complete laser relocalization and obtain the coarse pose of relocalization. Based on the coarse repositioning pose, a local map is extracted from the prior map. Point cloud registration is then performed between the current environment point cloud and the local map to obtain the fine repositioning pose, specifically including: The laser point clouds corresponding to the laser point cloud keyframes whose coarse poses for distance relocalization are within a preset range are combined together and downsampled to obtain a local map. A registration algorithm is used to register the currently observed environmental point cloud with the local map to obtain the fine pose for relocalization. When performing point cloud registration, the coarse pose for relocalization is used as the initial pose of the current environmental point cloud. The prior map includes a pre-constructed visual map and a laser map.

2. The mobile robot relocation method according to claim 1, characterized in that, The environmental image acquired by the visual sensor is matched with a prior map to complete visual relocalization and obtain a coarse pose for relocalization. Specifically: Extract the bag-of-words vector from the environmental image; The bag-of-words vector is compared with each bag-of-words vector in the bag-of-words database in the visual map, the bag-of-words vector in the bag-of-words database that is closest to the bag-of-words vector is found, and the image corresponding to the closest bag-of-words vector is obtained. The visual keyframes of the image are used as the coarse pose for relocalization.

3. The mobile robot relocation method according to claim 2, characterized in that, The step of comparing the bag-of-words vector with each bag-of-words vector in the bag-of-words database of the visual map, and finding the bag-of-words vector in the database that is closest to the bag-of-words vector, specifically involves: Calculate the similarity distance between the bag-of-words vectors and each bag-of-words vector in the bag-of-words database in the visual map; Select the bag-of-words vector corresponding to the minimum value from the obtained similarity distances, and take this bag-of-words vector as the closest bag-of-words vector.

4. The mobile robot relocation method according to claim 1, characterized in that, The environmental point cloud acquired by the laser sensor is matched with a prior map to complete laser relocalization and obtain a coarse pose for relocalization. Specifically: Extract the scan context descriptor from the environmental point cloud; The scan context descriptor is compared with each scan context descriptor in the scan context database of the laser map. The scan context descriptor that is closest to the scan context descriptor in the scan context database is found, and the laser point cloud corresponding to the closest scan context descriptor is obtained. The laser keyframes of the laser point cloud are used as the coarse pose for relocalization.

5. The mobile robot relocation method according to claim 4, characterized in that, The step of comparing the scan context descriptor with each scan context descriptor in the laser map scan context database, and finding the scan context descriptor in the scan context database that is closest to the scan context descriptor, specifically involves: Extract sub-descriptors from each scan context descriptor in the scan context database, and find candidate matching frames based on the sub-descriptors; Calculate the similarity distance between the scan context descriptor and the scan context descriptors of each candidate matching frame; Find the scan context descriptor of the candidate matching frame that is closest to the scan context descriptor. If the similarity distance between the two is less than a threshold, then the scan context descriptor of the closest candidate matching frame is taken as the closest scan context descriptor.

6. A mobile robot repositioning device, characterized in that, include: The visual relocalization module is used to match the environmental image acquired by the visual sensor with the prior map to complete the visual relocalization and obtain the coarse pose of the relocalization. The laser repositioning module is used to match the environmental point cloud collected by the laser sensor with the prior map when the visual repositioning module fails to reposition, and to complete the laser repositioning and obtain the coarse pose of the repositioning. A point cloud registration module is used to extract a local map from the prior map based on the coarse relocalization pose, and perform point cloud registration between the current environment point cloud and the local map to obtain the fine relocalization pose; the point cloud registration module includes: The local map generation unit is used to combine the laser point clouds corresponding to the laser point cloud keyframes with coarse poses of distance relocalization within a preset range, and perform downsampling operations to obtain a local map. The point cloud registration unit uses a registration algorithm to register the currently observed environmental point cloud with the local map to obtain the fine pose for relocalization. When performing point cloud registration, the coarse pose for relocalization is used as the initial pose of the current environmental point cloud. The prior map includes a pre-constructed visual map and a laser map.

7. The mobile robot repositioning device according to claim 6, characterized in that, The visual relocalization module includes: A bag-of-words vector extraction unit is used to extract the bag-of-words vectors of the environmental image; The bag-of-words vector comparison unit is used to compare the bag-of-words vector with each bag-of-words vector in the bag-of-words database in the visual map, find the bag-of-words vector in the bag-of-words database that is closest to the bag-of-words vector, and obtain the image corresponding to the closest bag-of-words vector. The first determining unit is used to use the visual image keyframes of the image as the coarse pose for relocalization.

8. The mobile robot repositioning device according to claim 6, characterized in that, The laser repositioning module includes: The descriptor extraction unit is used to extract the scan context descriptor of the environmental point cloud; The descriptor comparison unit is used to compare the scan context descriptor with each scan context descriptor in the scan context database of the laser map, find the scan context descriptor in the scan context database that is closest to the scan context descriptor, and obtain the laser point cloud corresponding to the closest scan context descriptor. The second determining unit is used to use the laser keyframe of the laser point cloud as the coarse pose for relocation.

9. A mobile robot, characterized in that, The system includes a memory and a processor, wherein the memory stores computer-executable instructions; and the processor executes the computer-executable instructions stored in the memory, causing the mobile robot to perform the mobile robot relocation method as described in any one of claims 1-6.

Images