Offline map fusion algorithm for indoor mobile robot based on line features

This study mainly investigates and analyzes the linear extraction of the center-line features of indoor mobile robots and point cloud maps. Based on the actual demand analysis of indoor mobile robots, an offline multi-map fusion algorithm based on improved ICP and line features was proposed. In general, the working environment of indoor mobile robots is constantly changing, and the robot does not know the fixed environment, which is the changing dynamic obstacle in the built map. Based on this problem, the algorithm can analyze the distribution of different obstacles in the same working environment for multiple robots. A robot builds multiple raster maps at different times and then proposes an offline multi-map fusion algorithm based on line features and improved ICP. Line features are extracted from maps, and the initial value Matrix between maps is calculated using points and line features matching to achieve rough matching between multiple maps. The ICP algorithm has high requirements for the initial values. The ICP algorithm can be used for fine matching. Finally, according to the map fusion results, only the repeated raster data in multiple maps were saved, and the weight of the unchanged raster points in the map was increased to improve the importance of the unchanged environment in the working process of the robot.