Path Selection for Mobile Robots in Dynamic Environments

Abstract

Abstract. This paper evaluates a path selection algorithm for mobile robots in large dynamic environments. The aim of the work is to reduce the risk of collisions and time of path following in cases when the robot repeatedly traverses between predefined target points (e.g. fot transportation or inspection tasks). The algorithm is usable even if very little is known about the environment or if it gets completely restructured during the mission. This article concentrates on evaluating the performance of the algorithm. The tests are completed on the mini-robot Khepera that operates in an unknown dynamically changing environment. The test results show that the path selection algorithm is able to reduce collision risk, travel time and travel distances as well as increase the predictability of robot’s behaviour and its degree of autonomy. 1