A correction system of odometry error for simultaneous map building based on sensor fusion

Abstract

This paper represents a SLAM(simultaneous localization and map building) system with odometry and indoor GPS sensor. This paper presents an accurate localization method by relative and absolute sensor fusion for map building system of mobile robot, we developed a sensor based absolution position tracking algorithm and grid based map building algorithm using this algorithm. A wall following decision engine like a fuzzy algorithm was proposed for map building navigation. Generally, while the robot is navigating the environment for map building position is estimated by odometry. However the position estimation using encoder data is not accurate owing to slipping and odometry linearization error. To solve this problem we proposed a sensor fusion method using encoder and indoor GPS. We proved this system' s validity through field test.