A Study on the Improved Localization of a Mobile Robot by using Scale Invariant Feature and Odometry

One of the important factors of the autonomous mobile robot is to build a map for surround environment and estimate its localization. This paper deals with the method to improve localization result by using scale invariant feature with single camera. In detail, the mobile robot attaches monocular camera looking wall extracts scale invariant features in each image by using SIFT (scale invariant feature transform). And then, map building is carried out by the extracted features and stores it map database. After finishing feature map building, the robot finds some points matched with previous feature map and its pose by affine parameter in real time. Localization by using SIFT is accurate relatively, but it sometimes loses its pose because of the change of the lights. To compensate the following problem, we apply the compensation method that uses the odometry information of the mobile robot. We verify the improved localization performance of the mobile robot from the experiments.