A numerical approach for planning robot trajectories considering collision avoidance

This paper describes a new approach for robot trajectory planning. Based on a numerical method for the inverse kinematic transformation, robot trajectories are planned by using a numerical optimization algorithm. The error between two tool center points (TCPs), a target TCP and a current search TCP is minimized by this algorithm. Collisions with static obstacles are avoided by using a potential field approach. An obstacle is wrapped with a potential field, figured by a Gaussian curve. The developed method is tested with simulation studies for the six-degree-of-freedom Puma 560 robot and then extended to a real Puma 560 robot