Computer vision capabilities for a semi-autonomous wheelchair

Abstract

The semi-autonomous wheelchair uses a multi-layered intelligent agent architecture designed to provide people with severe multiple disabilities navigational assistance in indoor environments. This paper describes the addition of computer vision capabilities to the semi-autonomous wheelchair. Specifically an intelligent agent design is described which detects doorways in indoor environments and provides the location to the user via a tactile interface. A prototype is developed that demonstrates the software architecture is capable of operating in a real time environment.