A recent development in robotics is the increase of intelligence in robots. One of the research fields is to enable robots to autonomously avoid collisions with surrounding objects. This article presents an efficient method for planning collision-free paths for an articulated robot that is surrounded by polyhedral objects. The algorithm plans a hypothetical Archimedes's spiral path from the initial position to the goal position. When a collision among the arms and obstacles is detected, the hypothetical path will be modified to avoid the collision. The algorithm applies geometric methods to determine the upper and lower bounds of the reachable area of the wrist and then determines a collision-free path point on that reachable area. Because the equations, which represent the upper and lower bounds, are simple, the algorithm can rapidly determine a collision-free path. Moreover, with minor modifications, this path planning algorithm can also be applied to other robots such as spherical, cylindrical, and Cartesian types of robots. (C) 1995 John Wiley & Sons, Inc.
