Vision Based Control for Power Assist Motion of Wheelchair Robots

This paper presents a vision based self-velocity estimation and its feedback system under force/torque sensor-less power assisting control of robotic wheelchair. In this method, three dimensional information obtained by stereo images, and the optical flow vectors are also used for self-velocity estimation in real-time. The human force is estimated by sensor-less reaction force observer, and the assist motion is generated by using its estimated force and virtual impedance model. In the paper, the force based assist function is integrated into visual feedback motion controller. This approach using vision and force based assist control makes it possible to facilitate the direct intelligent interactions between human force and environments such as human following assist, obstacle avoidance and so on. Such assist functions are changeable by the selection of the weighting matrix in the velocity estimation, which is based on weighted least square solutions from optical flow vectors. The validity of the proposed approach is verified by several experimental results