A collaborative wheelchair system.

This paper describes a novel robotic wheelchair, and reports experiments to evaluate its efficiency and understand how human operators use it. The concept at the heart of the collaborative wheelchair assistant (CWA) is to rely on the user's motion planning skills while assisting the maneuvering with flexible path guidance. The user decides where to go and controls the speed (including start and stop), while the system guides the wheelchair along software-defined guide paths. An intuitive path editor allows the user to avoid dangers or obstacles online and to modify the guide paths at will. By using the human sensory and planning systems, no complex sensor processing or artificial decision system is needed, making the system safe, simple, and low-cost. We investigated the performance of the CWA on its interaction with able-bodied subjects and motion efficiency. The results show that path guidance drastically simplifies the control. Using the CWA, the wheelchair user needs little effort from the first trial, while moving efficiently with a conventional wheelchair requires adaptation.