Outdoor navigation of a mobile robot with multiple sensors

Abstract

Outdoor navigation poses a challenge since uneven road surface, sunshine, clutter background will cause problems for odometry, laser, and vision sensors. To improve position accuracy, single strip retroreflective beacons have been used in the localization process. But to match observed beacons during the motion of the robot is a problem since all beacons are identical. Also strong reflective objects in outdoors may cause false reading.In this paper, we describe how we use the extended Kalman filter algorithm to integrate data scanned from the laser scanner rotating at 2Hz and readings from other sensors. The results obtained from outdoor navigation are presented.