Robust self-localization of ground vehicles using artificial landmark

摘要

Stable and accuracy localization has crucial importance on the success of autonomous navigation of robotic ground vehicles. In this paper, we present a landmark-based sensor fusion localization method that utilizes artificial landmarks, a cheap GPS sensor, and wheel odometry. We describe overall architecture of the proposed localization system and technical details of landmark-based localization algorithm. The developed landmark-based localization system gives centimeters localization accuracy and real time performance more than 100 Hz. The experimental results on extensive real video sequences confirms the effectiveness and reliability of the method, showing successful detection of most landmarks with almost no false detection.