A fast algorithm for moving obstacle avoidance for vision-based mobile robots

Abstract

In this paper, we present a geometric transformation-based motion planning algorithm for moving obstacle avoidance for mobile robots using artificial potential field. We show that this algorithm can be easily incorporated into various motion planning algorithms to achieve moving obstacles avoidance for kinematic/dynamic models of omnidirectional or nonholonomic robots. We have verified the proposed algorithm on a vision-based Hilare-type robot. Experiment results show that the proposed algorithm can effectively avoid unmodeled obstacles. Finally we prove that all derived control laws are stable using Lyapunov theory