A fuzzy logic approach to robotic path planning with obstacle avoidance

Abstract

Robotic path planning with obstacle avoidance is considered. The main objective of the proposed approach is to reduce planning time considerably for "real-time" applications with obstacles of arbitrary configuration. The concept of a minimum required joint motion envelope is introduced to facilitate the representation of the joint kinematic solutions. Near minimum time trajectories are generated through a phase plane analysis. If the planned path is obstructed, then the obstacle geometry is approximated and a suitable projection is found in joint space. A modified path is finally determined to avoid the obstacles. Fuzzy logic is employed to incorporate uncertainty into the algorithmic procedure. Preliminary investigations have been carried out using a PUMA 560 and its associated workspace.