Transformer substationinspection robot mapping method based on graph optimization

Abstract

The invention belongs to the technical field of intelligent robots, and discloses a transformer substation inspection robot mapping method based on graph optimization. The method comprises the following steps: obtaining an estimated value of a current pose of a robot by using an ICP algorithm; generating a subgraph by using laser data of a continuous specified number of frames, wherein the lattersubgraph comprises laser point cloud data of the specified number of frames in the previous subgraph; comparing the corresponding pose information of each frame of laser data in the subgraph with a key frame; when the pose information is similar to the key frame, performing closed loop connection thereon; and updating the robot poses in the subgraphs through a graph optimization method to form a final mapping result. By adoption of the transformer substation inspection robot mapping method based on graph optimization disclosed by the invention, the mapping efficiency of the robot is greatly improved, and the cumulative errors of the system are effectively reduced.

Description

technical field [0001] The invention belongs to the technical field of intelligent robots, and in particular relates to a map-building method for a substation inspection robot based on map optimization. Background technique [0002] Substation mapping is a necessary link before the operation of the laser navigation substation robot. It mainly records the map of the substation in the form of data as detailed and accurate as possible. The laser navigation substation inspection robot uses the map to compare with the surrounding information obtained by the laser sensor in real time, and determines the position of the substation inspection robot through a positioning algorithm. Therefore, the accuracy of the substation map is very important for the navigation and positioning of the laser navigation substation robot. [0003] Since SLAM (Simultaneous Localization And Mapping, real-time positioning and map construction) technology can well solve the problems of positioning, naviga...

AI Technical Summary

Problems solved by technology

[0006] (1) In the process of laser point cloud registration, due to the discrete nature of the laser point cloud data measured by the laser rangefinder, there is no very strict true relationship between the point cloud data of the same obstacle in two consecutive laser images match so that the laser data cannot effectively overlap

[0007] (2) Relying on the method of point cloud registration to build maps, there is a large cumulative error, and the established map is far from the real state

Images