Path mapping and planning with partially known paths using hierarchical state machine for service robot

Abstract

Path mapping is a very essential part of a mobile robot navigation system. In this work, a novel technique to map and plan path for a mobile service robot without any vision aids in indoor environment using hierarchical state machine with partially known paths is proposed. The known paths are taught to a robot using Learning by Demonstration technique (LfD). The first phase of the algorithm is to map the paths as a hierarchical state machine using the partially known paths. Second phase is to plan the path given the source and destination. The algorithm is implemented and tested using a 2D simulation environment platform, Player/Stage.