A probabilistic approach to collision-free robot path planning

A method was developed that takes into consideration the size and shape of the obstacles as well as the distance from the robot´s current position to the obstacles. The workspace is divided into a finite number of cubes. An obstacle is represented by the set of cubes it occupies. The probability of each cube becoming a dead end is calculated under the assumption that both the starting and destination points are randomly located in the workspace with a known distribution. These probabilities are used to determine an optimal robot path that minimizes some cost associated with the path. Several illustrative examples are presented. The method is intended to be used for 3-D problems and is simple enough to be implemented for online robot path planning