Evaluation of Collision Avoidance Path Planning Algorithm

This paper describes and evaluates a Collision Avoidance path planning Algorithm (CAA) for manipulator based on Minimum Distance Technique (MDT) that calculates collision free path by searching grid points in Euclidean Space. The key mechanism for avoiding collision based on MDT is the one that determine the minimum distance between two links and between a link and the edge of obstacles. It is necessary to compute the minimum distance between one link segment and another link segment (or edge of obstacles) when a link of manipulator is approaching collide with another link, manipulators and edge of obstacles. The basic idea of CAA is to solve collision avoidance problem for Manipulator between link-to-link, manipulator-to-manipulator, and manipulator-to-obstacles based on MDT. The novelty of CAA is that it calculate the shortest path from star position to goal position for manipulator with many links in fully complicated and cluttered environments. The performance of CAA is evaluated for 2-dimensional environments while changing the number of links and obstacles placement. The computational time also has been examined. Evaluation results show that CAA calculates the shortest path in relatively short period of the computation time and moreover, it achieves the constant performance independent of environments