Differential GPS and odometry-based outdoor navigation of a mobile robot

This paper describes outdoor navigation for a mobile robot by using differential GPS (DGPS) and odometry in a campus walkway environment. The robot position is estimated by fusion of DGPS and odometry. The GPS receiver measures its position by radio waves from GPS satellites. The error of GPS measurement data increases near high buildings and trees because of multi-path and forward diffractions. Thus, it is necessary to pick up only accurate DGPS measurement data when the robot position is modified by fusing DGPS and odometry. In this paper, typical DGPS measurement data observed near high buildings and trees are reported. Then, the authors propose a novel position correction method by fusing GPS and odometry. Fusion of DGPS and odometry is realized using an extended Kalman filter framework. Moreover, outdoor navigation for a mobile robot is accomplished by using the proposed correction method.