A visual servoing approach for autonomous corridor following and doorway passing in a wheelchair

Navigating within an unknown indoor environment using an electric wheelchair is a challenging task, especially if the user suffers from severe disabilities. In order to reduce fatigability and increase autonomy, control architectures have to be designed that would assist users in wheelchair navigation. We present a framework for vision-based autonomous indoor navigation in an electric wheelchair capable of following corridors, and passing through open doorways using a single doorpost. Visual features extracted from cameras on board the wheelchair are used as inputs for image based controllers built-in the wheelchair. It has to be noted that no a-priori information is utilized except for the assumption that the wheelchair moves in a typical indoor environment while the system is coarsely calibrated. The designed control schemes have been implemented onto a robotized wheelchair and experimental results show the robust behaviour of the designed system. We devise a visual servoing approach for indoor wheelchair navigation.The fundamental tasks of corridor following and doorway passing are tackled.Results in simulation and on a robotic platform show the convergence and validity of the control system.This work is proposed as a first step in developing the concept of semi-autonomous assistive systems.