Path Planning in a Dynamic Environment

Abstract

Path planning is an important areain the control of autonomous mobile robots.Recentwork hasfocusedon aspectsreductionsin processing timethanthe memory requirements. A dynamicenvironment usesa lot of memoryand hencethe processing timeincreases too.Ourapproach is to reducethe processing timebythe use of apictorial approachto reduce the numberofdata used. In this paper, we present apath planning approach that operates in three steps. First, a construction of the visibility tree is performed. The following treatments are not performed on the original image but on the result tree whose elements are specific points of the environment linked by the relationship of visibility. We construct thereafter a visibility graph which one seeks the shortest path. This approach has a great interest because of its fast execution speed. The path search is extendedalso for the case where obstacles can move. Themoving obstacles may be other mobile robots whose trajectories and speeds are known initially. At the end, someapplicationsareprovided on solving similarproblem suchcivil aviation in order toguideplane avoiding collisions.