A novel approach for motion planning

Abstract

The feasible map approach (FMA) to certain motion planning problems in robotics is introduced. This approach is based on the use of a feasible map representation of a configuration space. When a feasible map is constructed, the topology of a configuration space is derived by shrink transformation. Path planning is done at two levels: first, to find a feasible path in a feasible map, and second, to transform the feasible path in the feasible map into the configuration space. When a feasible map is obtained, the feasible path can be obtained according to any starting and desired configurations, at once. A detailed FMA is given on a sphere world and the performance of the proposed algorithm is demonstrated by experiments. The results indicate that the FMA is efficient, useful and widely applicable.