WO2022142744A1

WO2022142744A1 - Loopback detection method, apparatus and device, and computer readable storage medium

Info

Publication number: WO2022142744A1
Application number: PCT/CN2021/129493
Authority: WO
Inventors: 宁越; 张一凡; 邹李兵; 王学强
Original assignee: 歌尔股份有限公司
Priority date: 2021-01-04
Filing date: 2021-11-09
Publication date: 2022-07-07
Also published as: CN112700497B; CN112700497A

Abstract

Disclosed in the present invention are a loopback detection method, apparatus and device, and a computer readable storage medium. The method comprises: creating a multi-level to-be-compared local sub-graph, sequentially performing occupancy state comparison on the multi-level to-be-compared local sub-graph and a corresponding constructed local sub-graph according to levels, and obtaining the occupancy state matching degree of each level; if the occupancy state matching degree of each level is greater than the preset matching degree, performing occupancy state matching on a local sub-graph which is being constructed and the corresponding constructed local sub-graph; and if the occupancy states of the local sub-graph which is being constructed and the corresponding constructed local sub-graph at the same position are completely matched, determining that loopback succeeds. Therefore, after the occupancy states of the multi-level to-be-compared local sub-graphs are successfully matched, the local sub-graph which is being constructed and the corresponding constructed local sub-graph are subjected to state matching, and the loopback accuracy and the map construction success rate are improved.

Description

Loop closure detection method, apparatus, device, and computer-readable storage medium

technical field

The present invention relates to the technical field of image processing, and in particular, to a loop closure detection method, apparatus, device, and computer-readable storage medium.

Background of the Invention

When a 2D (two-dimensional, two-dimensional) laser is used to draw a map based on the SLAM (Simultaneous Localization and Mapping) algorithm, the map error will gradually accumulate and transmit, and a large cumulative error will occur for large maps.

The current main way to reduce the cumulative error is scan to map (matching the laser return data with the map), that is, directly matching the lidar scan data with the map to obtain the actual position coordinates. While calculating the position, the newly scanned data is added to the constructed map in time, but this method may loop back anywhere when the characteristic laser return data is relatively small, resulting in an inaccurate map.

SUMMARY OF THE INVENTION

Embodiments of the present invention provide a loopback detection method, apparatus, device, and computer-readable storage medium, which aim to improve the accuracy of loopback and the success rate of map construction.

To achieve the above purpose, an embodiment of the present invention provides a loopback detection method, which includes:

Create multi-level local sub-maps to be compared, and compare the occupancy states of the multi-level local sub-maps to be compared with the corresponding constructed local sub-maps in sequence to obtain the occupancy state matching degree of each level;

If the occupancy state matching degree of each level is greater than the preset matching degree, the occupancy state matching is performed between the local subgraph being constructed and the corresponding constructed local subgraph;

If the local subgraph being constructed completely matches the occupancy state of the corresponding constructed local subgraph at the same position, it is determined that the loopback is successful.

In addition, in order to achieve the above purpose, an embodiment of the present invention further provides a loopback detection device, where the loopback detection device includes:

The creation module is used to create a multi-level local sub-map to be compared, and compare the occupancy status of the multi-level local sub-map to be compared with the corresponding constructed local sub-map in turn to obtain the matching degree of occupancy status of each level;

The comparison module is used to match the occupancy state of the local sub-graph being constructed and the corresponding constructed local sub-graph if the occupancy state matching degree of each level is greater than the preset matching degree;

The determining module is configured to determine that the loopback is successful if the local sub-graph being constructed completely matches the occupancy state of the corresponding constructed local sub-graph at the same position.

In addition, in order to achieve the above object, an embodiment of the present invention also provides a loopback detection device, the loopback detection device includes a processor, a memory, and a loopback detection program stored in the memory. When the loopback detection program is run by the processor, the above-mentioned Steps of a loopback detection method.

In addition, in order to achieve the above object, an embodiment of the present invention further provides a computer-readable storage medium storing a loopback detection program on the computer-readable storage medium. The loopback detection program is executed by a processor to implement the steps of the above loopback detection method.

Compared with the prior art, a method, device, device and computer-readable storage medium for loop closure detection proposed by the present invention create multi-level partial sub-maps to be compared, and sequentially associate the multi-level partial sub-maps to be compared with corresponding ones according to the level. The occupancy states of the constructed local subgraphs are compared to obtain the occupancy state matching degree of each level; if the occupancy state matching degree of each level is greater than the preset matching degree, the local subgraph being constructed is compared with the corresponding constructed local subgraph. The graph performs occupancy state matching; if the local subgraph being constructed completely matches the occupancy state at the same position of the corresponding constructed local subgraph, it is determined that the loopback is successful. Therefore, after the occupancy state of the multi-level to-be-compared partial subgraph is successfully matched, the state of the local subgraph being constructed is matched with the corresponding constructed local subgraph, which improves the accuracy of loop closure and the success rate of map construction.

Brief Description of Drawings

1 is a schematic diagram of a hardware structure of a loopback detection device involved in various embodiments of the present invention;

FIG. 2 is a schematic flowchart of the first embodiment of the loopback detection method of the present invention;

FIG. 3 is a schematic diagram of functional modules of the first embodiment of the loopback detection apparatus of the present invention.

The realization, functional characteristics and advantages of the present invention will be further described with reference to the accompanying drawings in conjunction with the embodiments.

Detailed ways

It should be understood that the specific embodiments described herein are only used to explain the present invention, but not to limit the present invention.

The loopback detection device mainly involved in the embodiments of the present invention refers to a network connection device capable of realizing network connection, and the loopback detection device may be a server, a cloud platform, or the like.

Referring to FIG. 1 , FIG. 1 is a schematic diagram of a hardware structure of a loopback detection device involved in various embodiments of the present invention. In this embodiment of the present invention, the loopback detection device may include a processor 1001 (for example, a central processing unit, Central Processing Unit, CPU), a communication bus 1002, an input port 1003, an output port 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection communication between these components; the input port 1003 is used for data input; the output port 1004 is used for data output, and the memory 1005 can be a high-speed RAM memory or a non-volatile memory (non-volatile memory). memory), such as a disk memory, the memory 1005 may optionally also be a storage device independent of the aforementioned processor 1001 . Those skilled in the art can understand that the hardware structure shown in FIG. 1 does not constitute a limitation of the present invention, and may include more or less components than those shown in the drawings, or combine some components, or arrange different components.

Continuing to refer to FIG. 1 , the memory 1005 in FIG. 1 as a readable computer-readable storage medium may include an operating system, a network communication module, an application program module, and a loopback detection program. In FIG. 1 , the network communication module is mainly used to connect to the server and perform data communication with the server; and the processor 1001 can call the loopback detection program stored in the memory 1005 and execute the loopback detection method provided by the embodiment of the present invention.

Embodiments of the present invention provide a loopback detection method.

Referring to FIG. 2, FIG. 2 is a schematic flowchart of the first embodiment of the loopback detection method of the present invention.

In this embodiment, a loopback detection method is applied to a loopback detection device, and the method includes:

Step S101, creating a multi-level local subgraph to be compared, and comparing the occupancy state of the multi-level local subgraph to be compared with the corresponding constructed local subgraph in turn according to the level, to obtain the occupancy state matching degree of each level;

Loopback detection, also known as closed-loop detection, refers to the ability of the robot to recognize that it has reached a certain scene and make the map closed-loop. To put it simply, when the robot turns left and right to build a map, it can realize that a certain place has come before, and then match the map generated at this moment with the previously generated map.

This embodiment draws a map based on SLAM (Simultaneous Localization and Mapping). SLAM is mainly used to solve the problems of positioning, navigation and map construction when mobile robots are running in unknown environments. The process of building a local subgraph through SLAM includes:

Create a blank subgraph centered on the robot's location;

The robot is equipped with a laser sensor, which can emit laser light and receive laser return data to measure the distance from the obstacle to the excitation sensor. For example, the laser sensor can be a 2D laser sensor. After the 2D laser sensor emits a frame of laser light, it receives the return laser return data in all directions, and determines the distance of the obstacle according to the return time. Based on the time difference from the launch of the laser to the receipt of the returned laser data, multiply the speed and divide by two to get the distance from the sensor to the nearest obstacle in the corresponding direction.

During the process of building the map, the robot is constantly moving and can fire lasers at various locations and receive return data. In this embodiment, a blank subgraph is created centered on the location where the robot emits laser light and receives returned data. The size of the blank sub-image can be determined based on the actual scene.

Convert the laser return data collected by the robot to a pre-built blank sub-map, mark the occupancy status of each position in the blank sub-map, and obtain a partial sub-map; splicing multiple continuous partial sub-maps into a constructed partial sub-map .

In this embodiment, the coordinate system where the robot and its laser return data are located is marked as the world coordinate system, the collected laser return data is marked as scan points in the world coordinate system, the location of the robot is marked as origin, and the robot is marked as origin. The location and laser return data are expressed as H={origin, scan points}. Create a map coordinate system in the blank subgraph, and denote the transformation matrix between the world coordinate system and the map coordinate system as Tξ, so that the position of the robot and the laser return data H can be converted to the blank subgraph through Tξ. The occupancy state at the position corresponding to the coordinates of the obstacle is marked as occupied in the blank subgraph.

In this embodiment, the blank submap and the converted partial submap (submap) can be a grid map, the size of which can be 5cm×5cm, and the center of each partial submap is called a grid point (grid point). , and other points in the local subgraph are called corresponding pixels. By splicing multiple consecutive local subgraphs based on the SLAM algorithm, the final result is a constructed local map containing many subgraphs.

The local submap can be regarded as a grid probability map, but if the local subgraph is regarded as a feature point, then the entire constructed local map can be regarded as a topological map, and the local subgraphs are the feature points, and the relative positions between the local subgraphs The constraint relationship is a line. The relationship between the local subgraph and the local subgraph is maintained through the relative position of the center point, but the relative position error is relatively low between the local subgraphs with similar positions. Various random factors, etc., the error will become larger and larger. For example, assuming that there is an error of 1em in the adjacent local subgraph, when the final constructed map contains 100 local subgraphs, the last local subgraph is the same as the first one. The error between the local subgraphs is as much as 1m, so loopback detection is required to correct these errors.

In addition, when splicing the constructed local map, PL-ICP (piont line-iterative closest point, point-line iteration closest point) is used to correct the accumulated error caused by the robot in the process of small-scale motion, and then according to the corrected robot pose , which converts the data based on the new robot pose and its laser return into a local subgraph. When the laser return data stored in the local sub-map reaches the preset number, the construction of the local sub-map can be completed.

In this embodiment, the multi-level local sub-images to be compared refer to local sub-images to be compared with different resolutions.

Create multi-level partial subgraphs to be compared, including:

Compress the constructed local sub-images at a preset resolution to obtain multiple hierarchically compressed local sub-images of different resolutions. The preset resolutions include a first preset resolution, a second preset resolution, and a third preset resolution. Resolution, the corresponding hierarchical compression local sub-images are the first-level compressed local sub-image, the second-level compressed local sub-image and the third-level compressed local sub-image; the laser signal data received by the robot are respectively converted to the hierarchical compression local sub-image. Subgraph, obtain the local subgraph to be compared. In this embodiment, based on the SLAM algorithm, the laser signal data are respectively converted into hierarchically compressed local sub-maps.

Generally, the resolution of the compressed hierarchically compressed local sub-map is lower than the corresponding constructed local sub-map. After the high-resolution constructed local sub-image is compressed, the resolution is reduced and the number of pixels is reduced. For example, if the area of the constructed local sub-image A is 1 square meter, when the resolution is 5cm per pixel, there are 400 pixels; if the compression is 50cm per pixel, there are 4 pixels. For the constructed local sub-images of the same size, the number of pixels is reduced, and the amount of computation required for matching is greatly reduced. For example, three layers can be set, and each layer corresponds to the first preset resolution, the second preset resolution, and the third preset resolution. For example, the first preset resolution is 5cm per pixel, and the second preset resolution is The rate is 10cm per pixel and the third preset resolution is 50cm.

Specifically, the first occupancy state of each position in the first-level compressed local submap is compared with the occupancy state to be compared at each corresponding position in the corresponding constructed local submap to obtain the first occupancy state matching degree; this embodiment The resolution of the first-level compressed local sub-image is the first resolution. The first occupancy status of the corresponding positions of each pixel point in the first-level compressed local sub-map is acquired, and at the same time, the to-be-compared occupancy status corresponding to each pixel point position in the constructed local sub-map is acquired. The first occupancy state of each position is compared with the occupancy state to be compared to obtain the matching degree of the first occupancy state. The resolution of the first-level compressed local sub-image is low, the pixel points are relatively small, there are not many positions to be compared, and the amount of calculation is small, so that the local sub-images with obvious differences can be screened.

If the matching degree of the first occupancy state is greater than the preset first matching degree, the second occupancy state of each position in the second-level compressed local submap is compared with the occupancy state of each corresponding position in the corresponding constructed local submap to be compared. By comparison, a second occupancy state matching degree is obtained, and the second preset resolution is higher than the first preset resolution; the preset first matching degree can be specifically set as required, for example, the preset first matching degree is set to 80%, 85% or 90%. Since the resolution of the second-level compressed local sub-image is higher than that of the first-level compressed local sub-image, it has more pixels and can perform more accurate matching.

If the matching degree of the second occupancy state is greater than the preset second matching degree, the third occupancy state of each position in the third-level compressed local submap is compared with the occupancy state of each corresponding position in the corresponding constructed local submap to be compared. By comparison, a third occupancy state matching degree is obtained, and the third preset resolution is higher than the second preset resolution. The preset second matching degree can be specifically set as required, for example, the preset second matching degree is set to 85% or 90%.

In addition, if the first occupancy state matching degree is less than or equal to the preset first matching degree, performing error correction and updating on the constructed local sub-map based on the laser signal data received by the robot;

If the matching degree of the second occupancy state is less than or equal to the preset second matching degree, error correction and updating of the constructed local sub-map is performed based on the laser signal data received by the robot.

If the matching degree of the first occupancy state and/or the degree of matching of the second occupancy state is low, it indicates that there may be errors in the constructed local sub-images, and therefore it is necessary to perform error correction and update in time to ensure the accuracy of the composition.

Step S102, if the occupancy state matching degree of each level is greater than the preset matching degree, the occupancy state matching is performed between the local subgraph being constructed and the corresponding constructed local subgraph;

If the first occupancy state matching degree, the second occupancy state matching degree and the third occupancy state matching degree are all greater than the corresponding preset matching degrees, it means that the similarity between the local subgraph being constructed and the corresponding constructed local subgraph is relatively high. If the value is high, it means that the correctness of the constructed local subgraph is relatively high, that is, the laser return signal is successfully matched, and further loopback judgment is performed.

Specifically, the current occupancy status of each position in the local submap being constructed is matched with the occupancy status to be compared at each corresponding position in the corresponding constructed local submap.

Step S103, if the local subgraph being constructed completely matches the occupancy state of the corresponding constructed local subgraph at the same position, it is determined that the loopback is successful.

Place the local sub-image being constructed to the position where the laser matching was successful before, and after coordinate transformation, if the local sub-image being constructed and the local sub-image already constructed on this match have the same occupancy status at the same coordinates, then The loopback is considered successful.

If the loopback detection is successful, the accumulated error can be significantly reduced, helping the robot to perform obstacle avoidance and navigation more accurately and quickly. And wrong detection results can make the map bad. Therefore, loop closure detection is very necessary in the construction of large-area and large-scene maps.

After the laser matching is successful, the local submap being constructed is compared with the corresponding constructed local submap, so the loop closure detection method in this application is actually a map_to_map (map to map) method. The map_to_map loopback detection method is to use not only a single frame of laser but also an unfinished local submap submap for each loopback detection. It is very easy to mismatch in the environment of , but the local submap contains map information for a period of time, and describes more environmental features, so it can greatly reduce the possibility of mismatch. After the laser matching is successful, the local subgraph is compared with the constructed subgraph, and a second judgment is made. If the matching is still determined, the loopback matching is considered successful. This will not only not affect the real-time performance of the algorithm, but also improve the reliability of the algorithm output. The front-end and back-end of the algorithm are generally operated by two processes at the same time. This solution can ensure the overall real-time performance of the algorithm.

Further, after step S103, the method further includes: optimizing the constructed local subgraphs to reduce the accumulated error of the constructed local subgraphs.

In this embodiment, optimization may be performed based on the error of the adjacent subgraphs in the constructed local subgraph to minimize the final error, and the constructed local subgraph with the smallest error is saved as the final constructed local subgraph.

In this embodiment, through the above solution, a multi-level local sub-map to be compared is created, and the multi-level local sub-map to be compared is sequentially compared with the corresponding constructed local sub-map according to the occupancy status, so as to obtain the matching degree of occupancy status of each level. ; If the matching degree of occupancy status of each level is greater than the preset matching degree, the occupancy status of the local subgraph being constructed is matched with the corresponding constructed local subgraph; if the local subgraph being constructed matches the corresponding constructed local subgraph If the occupancy states of the subgraphs at the same position completely match, it is determined that the loopback is successful. Therefore, after the occupancy state of the multi-level to-be-compared partial subgraph is successfully matched, the state of the local subgraph being constructed is matched with the corresponding constructed local subgraph, which improves the accuracy of loop closure and the success rate of map construction.

In addition, this embodiment also provides a loopback detection device. Referring to FIG. 3 , FIG. 3 is a schematic diagram of functional modules of the first embodiment of the loopback detection apparatus of the present invention.

In this embodiment, the loopback detection device is a virtual device, which is stored in the memory 1005 of the loopback detection device shown in Compare the occupancy states of the multi-level local sub-maps to be compared with the corresponding constructed local sub-maps in turn to obtain the occupancy state matching degree of each level; it is used if the occupancy state matching degree of each level is greater than the preset matching degree, then Match the occupancy state of the local subgraph under construction with the corresponding constructed local subgraph; it is used to determine that the loopback is successful if the local subgraph being constructed completely matches the occupancy state of the corresponding constructed local subgraph at the same position .

Specifically, the loopback detection device includes:

The creation module 10 is used to create a multi-level local sub-graph to be compared, and the multi-level local sub-graph to be compared is successively compared with the corresponding constructed local sub-graph by level, and the occupancy state matching degree of each level is obtained;

The comparison module 20 is used for matching the occupancy state of the local sub-graph being constructed and the corresponding constructed local sub-graph if the occupancy state matching degree of each level is greater than the preset matching degree;

The determination module 30 is configured to determine that the loopback is successful if the occupancy state of the local sub-graph being constructed and the corresponding constructed local sub-graph at the same position are completely matched.

Further, creating modules is also used to:

Compress the constructed local sub-images at a preset resolution to obtain multiple hierarchically compressed local sub-images of different resolutions. The preset resolutions include a first preset resolution, a second preset resolution, and a third preset resolution. Resolution, the corresponding hierarchical compressed local sub-images are the first-level compressed local sub-image, the second-level compressed local sub-image and the third-level compressed local sub-image;

The laser signal data received by the robot are respectively converted into hierarchical compressed local sub-images to obtain the local sub-images to be compared.

Further, creating modules is also used to:

Comparing the first occupancy status of each position in the first-level compressed local submap with the to-be-compared occupancy status of each corresponding position in the corresponding constructed local submap to obtain the first occupancy status matching degree;

If the matching degree of the first occupancy state is greater than the preset first matching degree, the second occupancy state of each position in the second-level compressed local submap is compared with the occupancy state of each corresponding position in the corresponding constructed local submap to be compared. By comparison, a second occupancy state matching degree is obtained, and the second preset resolution is higher than the first preset resolution;

If the matching degree of the second occupancy state is greater than the preset second matching degree, the third occupancy state of each position in the third-level compressed local submap is compared with the occupancy state of each corresponding position in the corresponding constructed local submap to be compared. By comparison, a third occupancy state matching degree is obtained, and the third preset resolution is higher than the second preset resolution.

Further, creating modules is also used to:

If the first occupancy state matching degree is less than or equal to the preset first matching degree, error correction and updating of the constructed local sub-map is performed based on the laser signal data received by the robot.

Further, creating modules is also used to:

Create a blank subgraph centered on the robot's location;

Convert the laser return data collected by the robot to a pre-built blank sub-map, mark the occupancy status of each position in the blank sub-map, and obtain a local sub-map;

Concatenates multiple consecutive local subgraphs into a constructed local subgraph.

Further, the comparison module is also used to:

The current occupancy status of each position in the local sub-map being constructed is matched with the to-be-compared occupancy status of each corresponding position in the corresponding constructed local sub-map.

Further, the comparison module is also used to:

The constructed local subgraph is optimized to reduce the accumulated error of the constructed local subgraph.

In addition, an embodiment of the present invention further provides a computer-readable storage medium, where a loopback detection program is stored on the computer-readable storage medium, and when the loopback detection program is run by a processor, the steps of the above loopback detection method are implemented.

It should be noted that, herein, the terms "comprising", "comprising" or any other variation thereof are intended to encompass non-exclusive inclusion, such that a process, method, article or system comprising a series of elements includes not only those elements, It also includes other elements not expressly listed or inherent to such a process, method, article or system. Without further limitation, an element qualified by the phrase "comprising a..." does not preclude the presence of additional identical elements in the process, method, article or system that includes the element.

The above-mentioned serial numbers of the embodiments of the present invention are only for description, and do not represent the advantages or disadvantages of the embodiments.

From the description of the above embodiments, those skilled in the art can clearly understand that the method of the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is better implementation. Based on such understanding, the technical solutions of the present invention can be embodied in the form of software products in essence or the parts that contribute to the prior art, and the computer software products are stored in the above-mentioned computer-readable storage medium (such as ROM/RAM, magnetic disk, optical disk), including several instructions to make a terminal device execute the methods described in the various embodiments of the present invention.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the scope of the present invention. Any equivalent structure or process transformation made by using the contents of the description and drawings of the present invention, or directly or indirectly applied in other related technical fields , are similarly included in the scope of patent protection of the present invention.

Claims

A loop closure detection method, wherein the method comprises:

Create a multi-level local sub-map to be compared, and compare the occupancy status of the multi-level local sub-map to be compared with the corresponding constructed local sub-map in turn according to the level, so as to obtain the occupancy status matching degree of each level;

If the occupancy state matching degree of each level is greater than the preset matching degree, the occupancy state matching is performed between the local subgraph being constructed and the corresponding constructed local subgraph;

The loopback is determined to be successful if the occupancy status of the local sub-graph being constructed and the corresponding constructed local sub-graph at the same position completely match.
The method according to claim 1, wherein the creating a multi-level partial subgraph to be compared comprises:

Compress the constructed local sub-images at a preset resolution to obtain multiple hierarchically compressed local sub-images of different resolutions, where the preset resolutions include a first preset resolution, a second preset resolution, and a third preset resolution. At the preset resolution, the corresponding hierarchical compressed local sub-images are respectively the first-level compressed local sub-image, the second-level compressed local sub-image, and the third-level compressed local sub-image;

The laser signal data received by the robot are respectively converted into the hierarchical compressed local sub-maps to obtain the local sub-maps to be compared.
The method according to claim 2, wherein the multi-level local sub-graphs to be compared are sequentially compared with the corresponding constructed local sub-graphs according to the occupancy state to obtain the occupancy state matching degree of each level, comprising: :

Comparing the first occupancy state of each position in the first-level compressed local submap with the to-be-compared occupancy state of each corresponding position in the corresponding constructed local submap, to obtain the first occupancy state matching degree;

If the matching degree of the first occupancy state is greater than the preset first matching degree, the second occupancy state of each position in the second-level compressed local sub-map and the corresponding position in the corresponding constructed local sub-map to be comparing the occupancy states to obtain a second occupancy state matching degree, where the second preset resolution is higher than the first preset resolution;

If the matching degree of the second occupancy state is greater than the preset second matching degree, the third occupancy state of each position in the third-level compressed local sub-map and the corresponding position in the corresponding constructed local sub-map to be The occupancy states are compared to obtain a third occupancy state matching degree, where the third preset resolution is higher than the second preset resolution.
The method according to claim 3, wherein the comparing the first occupancy status of each position in the first-level compressed local sub-map with the to-be-compared occupancy status of each corresponding position in the corresponding constructed local sub-map , obtain the first occupancy state matching degree, and then include:

If the first occupancy state matching degree is less than or equal to the preset first matching degree, performing error correction and updating on the constructed local sub-map based on the laser signal data received by the robot.
The method according to claim 1, wherein, creating a multi-level partial sub-map to be compared, and comparing the occupancy status of the multi-level partial sub-map to be compared with the corresponding constructed local sub-map in sequence, Obtain the occupancy status match of each level, which previously included:

Create a blank subgraph centered on the robot's location;

Convert the laser return data collected by the robot into a pre-built blank sub-map, mark the occupancy status of each position in the blank sub-map, and obtain a local sub-map;

A plurality of consecutive local subgraphs are stitched into the constructed local subgraph.
The method according to claim 1, wherein the matching of the occupancy state between the local sub-graph being constructed and the corresponding constructed local sub-graph comprises:

Matching the current occupancy status of each position in the local submap under construction with the occupancy status to be compared at each corresponding position in the corresponding constructed local submap.
The method according to any one of claims 1-6, wherein, if the local sub-graph being constructed completely matches the occupancy state at the same position of the corresponding constructed local sub-graph, it is determined that the loopback is successful, After that it also includes:

The constructed local subgraphs are optimized to reduce the accumulated error of the constructed local subgraphs.
A loopback detection device, wherein the loopback detection device comprises:

The creation module is used to create a multi-level local sub-map to be compared, and compare the occupancy status of the multi-level local sub-map to be compared with the corresponding constructed local sub-maps in turn to obtain the occupancy status matching degree of each level. ;

The comparison module is used to match the occupancy state of the local subgraph being constructed and the corresponding constructed local subgraph if the occupancy state matching degree of each level is greater than the preset matching degree;

A determination module, configured to determine that the loopback is successful if the occupancy state of the local sub-graph being constructed and the corresponding constructed local sub-graph at the same position are completely matched.
A loopback detection device, wherein the loopback detection device includes a processor, a memory, and a loopback detection program stored in the memory. When the loopback detection program is run by the processor, the following loopback detection method is implemented:

Create a multi-level local sub-map to be compared, and compare the occupancy status of the multi-level local sub-map to be compared with the corresponding constructed local sub-map in turn according to the level, so as to obtain the occupancy status matching degree of each level;

If the occupancy state matching degree of each level is greater than the preset matching degree, the occupancy state matching is performed between the local subgraph being constructed and the corresponding constructed local subgraph;

The loopback is determined to be successful if the occupancy status of the local sub-graph being constructed and the corresponding constructed local sub-graph at the same position completely match.
A computer-readable storage medium, wherein a loopback detection program is stored on the computer-readable storage medium, and when the loopback detection program is run by a processor, the loopback detection method according to any one of claims 1-7 is implemented A step of.
The apparatus of claim 8, wherein the creation module is further configured to:

Compress the constructed local sub-images at a preset resolution to obtain multiple hierarchically compressed local sub-images of different resolutions. The preset resolutions include a first preset resolution, a second preset resolution, and a third preset resolution. Resolution, the corresponding hierarchical compressed local sub-images are the first-level compressed local sub-image, the second-level compressed local sub-image and the third-level compressed local sub-image;

The laser signal data received by the robot are respectively converted into hierarchical compressed local sub-images to obtain the local sub-images to be compared.
The apparatus of claim 11, wherein the creation module is further configured to:

Comparing the first occupancy status of each position in the first-level compressed local submap with the to-be-compared occupancy status of each corresponding position in the corresponding constructed local submap to obtain the first occupancy status matching degree;

If the matching degree of the first occupancy state is greater than the preset first matching degree, the second occupancy state of each position in the second-level compressed local submap is compared with the occupancy state of each corresponding position in the corresponding constructed local submap to be compared. By comparison, a second occupancy state matching degree is obtained, and the second preset resolution is higher than the first preset resolution;

If the matching degree of the second occupancy state is greater than the preset second matching degree, the third occupancy state of each position in the third-level compressed local submap is compared with the occupancy state of each corresponding position in the corresponding constructed local submap to be compared. By comparison, a third occupancy state matching degree is obtained, and the third preset resolution is higher than the second preset resolution.
The apparatus of claim 8, wherein the creation module is further configured to:

Create a blank subgraph centered on the robot's location;

Convert the laser return data collected by the robot to a pre-built blank sub-map, mark the occupancy status of each position in the blank sub-map, and obtain a local sub-map;

Concatenates multiple consecutive local subgraphs into a constructed local subgraph.