A flexible algorithm for planning local shortest path of mobile robots based on reachability graph

Proposes concept called 'local shortest path' for mobile robots, and shows that a more compact V-graph of size O(M/sup 2/+N) can be constructed based on this concept, where M and N are numbers of convex components and convex vertices of polygonal obstacles respectively. In addition, a reachability graph (R-graph) of size O(N*M/sup 2/) registering reachability between vertices on local shortest paths is proposed. The R-graph depends only on obstacles in the environment but not on size of mobile robots. Hence even if the size of the robot or the required safety distance between the robot and obstacles changed, it is possible to plan a path for the robot efficiently by picking up its reachable vertices in the R-graph without need of reconstruction of the R-graph. Finally, the usefulness of the algorithm is ascertained by several simulations.<<ETX>>