BEHAVIOR-BASED MAP REPRESENTATION FOR A SONAR-BASED MOBILE ROBOT BY STATISTICAL-METHODS

Many conventional methods for map generation by mobile robots have tri ed to reconstruct a 3D geometric representation of the environment, wh ich is time-consuming, error-prone and necessary to transform the map into the information available for the given task. This paper proposes a method to acquire a statistical map representation robust to sensor noise and directly usable for a navigation task. The robot is equippe d with a ring of ultrasonic ranging sensors and a collision avoidance behavior is embedded in it. First, the mobile robot explores the envir onment in order to store a set of sequences of sonar data and principl e component analysis is applied to reduce the dimensionality of the so nar data. As a result, each sequence of sonar data can be described as a score pattern of principal components. Next, these patterns are cla ssified into typical local structures of the environment in order for the robot to discriminate them. Finally, a graph representation of the environment is constructed in which the nodes and arcs correspond to these local structures and the transition probabilities between them, respectively. The validity of the method is shown by computer simulati ons and real robot experiments.