3D Localization of a Mobile Robot by Using Monte Carlo Algorithm and 2D Features of 3D Point Cloud

Modern buildings are designed with wheelchair accessibility, giving an opportunity for wheeled robots to navigate through sloped areas while avoiding staircases. Nonetheless, working safely and autonomously in uneven or unstructured environments is still challenging for mobile robots. Herein, we propose the use of a point cloud treatment and Monte Carlo localization in an algorithm for 3D localization of a mobile robot in uneven and unstructured environments. The point cloud treatment consists of projecting the acquired 3D point cloud and generated 2D features. These features are used to recreate an occupancy map. To perform localization in real-time, the robot systems use the occupancy map and 2D features as input in conjunction with the Monte Carlo algorithm.