Research on Lidar Localization Based on Indoor Structural Feature Extraction

Mobile robots used in modern automation must have strong localization capabilities in challenging environments. This work suggests a method of lidar artificial landmark localization that exploits structural elements of buildings as artificial landmarks instead of reflectors in light of the drawbacks of typical indoor lidar artificial landmark localization, which necessitates reflectors. The map must first be created in order to determine the position of the building’s structural features before the structural components of the building can be extracted and stored in the specified file. In the navigation process, the structural elements of the building are extracted in real time and compared with the pre-stored position, and threshold values are set to eliminate abnormal recognition. It is crucial to confirm the discovered features since environmental changes and extracted features could be similar, leading to false positives. The proposed algorithm can improve the localization accuracy of mobile robots in complex dynamic environments compared with Cartographer pure localization, according to experimental results.