EKF and Computer Vision for Mobile Robot Localization

Abstract

The autonomous robotic system accurate localization is a challenging step in robot navigation field once the mobile device should avoid dangerous situations, such as unsafe conditions and collisions. In this context, the present paper proposes a localization method using the Extended Kalman Filter (EKF) to fuse the information coming from two different sensors (i.e. odometry and computer vision). The localization results present with known and unknown starting points and are tested in a simulated environment.