Mobile robot SLAM for line-based environment representation

This paper presents an algorithm for solving the simultaneous localization and map building (SLAM) problem, a key issue for autonomous navigation in unknown environments. The considered scenario is that of a mobile robot using range scans, provided by a 2D laser rangefinder, to update a map of the environment and simultaneously estimate its position and orientation within the map. The environment representation is based on linear features whose parameters are extracted from range scans, while the corresponding covariance matrices are computed from the statistical properties of the raw data. Simultaneous update of robot pose and linear feature estimates is performed via extended Kalman filtering. Experimental tests performed within a real-world indoor environment demonstrate the effectiveness of the proposed SLAM technique.