Abstract
The paper presents a new approach to the problem of map building and simultaneous localization (SLAM), originally developed by this research. The global map is built by continuously matching the feature polylines obtained from the laser ranger finder. The odometry error between each constitutive step can be eliminated by the global map, which was built in the previous step. The localization of the robot is done using the Monte Carlo approach. The experimental result is obtained by Home Robot, developed by Toshiba Corporation.
