A reduced cascaded fuzzy logic controller for dynamic window weights optimization

Abstract

Mobile robots in indoor environments require path planning techniques with very fast processing cycles and ability to work in dynamic environments. The dynamic window approach is one of the most sophisticated approaches in obstacle avoidance. It uses objective function to choose the optimal velocity commands. The objective function consist of three components. Changing the weights of these component will change the behavior of the algorithm. In this paper, a fuzzy logic controller that depends only on the obstacles distribution around the robot and uses less number of rules in the fuzzy logic is proposed. When compared with the other fuzzy logic controller this approach is much faster, reducing the processing cycle time for the overall approach and also the path obtained is smoother and with higher velocities in areas crowded with obstacles.