Heuristic motion planning with many degrees of freedom

Abstract

We present a general heuristic approach to the geometric motion planning problem with the aim to quickly solve intuitively simple problems. It is based on a divide-and-conquer path search strategy which makes inquiries about feasible paths; to answer these, we develop an efficient collision detection scheme that handles translations and rotations of polyhedra to compute all times of collision. The whole algorithm can be easily implemented and universally applied and has been successfully tested in a program for assembly planning.