Visual servoing with fiducial landmarks for differential drive mobile robot

This paper aims to present a visual servoing control strategy that uses visual information, acquired by a fiducial system, to solve the regulation to an arbitrary pose problem in differential driven mobile robots. In addition, in order to make this system more robust and immune to occlusion, we will use an Extended Kalman Filter to fuse odometric data and the data acquired from several landmarks. To demonstrate the performance of the approach, we present some experimental results.