An enhanced path planning of fast mobile robot based on data fusion of image sensor and GPS

Abstract

This paper presents a path planning algorithm for a fast mobile robot based on Extended Kalman Filter (EKF) by fusing the satellite navigation and the vision system in the outdoor environment. Lhe suggested approach offers several improvements that result in smoother trajectories and greater reliability. Lhe noisy location information of a robot is enhanced by using the vision system which contains abundant information with high accuracy but is subject to noise. Lhis research consists of a motion segmentation stage which gets motion information of moving objects form motion model, and a motion estimation stage which estimates the position and the motion of moving object using EKF. EKF based approach is served as the de-facto approach to SLAM with shortcomings of quadratic complexity and sensitivity to failures in data association. Lhe simulation results show a greater reliability for fast mobile robot navigation under outdoor environment. © 2009 SICE.