Vision-Based Shared Control for a BCI Wheelchair

Brain-actuated wheelchairs offer paraplegics the potential to gain a degree of independence in performing activities of daily living. It is not currently possible to achieve precise proportional control of devices using the low resolution output of a brain-computer interface (BCI). Consequently, we have developed a shared control system that interprets such commands, given the context of the surroundings. In this paper we show that a vision system provides sufficiently reliable information to the shared controller, to enable synthesized BCI subjects to drive safely in an office environment. The shared controller reduces both the time and number of commands required to perform a task.