Cellular robotic collision-free path planning

Abstract

Presents a massively parallel, connectionist model for cellular robotic collision-free path planning. A number of cellular robots (represented by via points) are distributed along the initial path arbitrarily set up, and move toward the final positions to form an optimal collision-free path, following the locally defined rules. The local rules for individual cellular robots are based on the gradient of the potential field due to the polyhedral obstacles and the distances to the neighboring cellular robots. The cellular robotic path planning provides massive parallelism and has power to effectively deal with path planning of three dimensional case with translational and rotational motions. Finally, the algorithm incorporates simulated annealing to solve a local minimum problem.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>