ABSTRACT
Many of the physically injured use electric wheelchairs as an aid to locomotion. Usually, for commanding this type of wheelchair, it is required the use of one’s hands and this poses a problem to those who, besides being unable to use their legs, are also unable to properly use their hands. The aim of the work described here, is to create a prototype of a wheelchair command interface that do not require hand usage. Facial expressions were chosen instead, to provide the necessary visual information for the interface to recognize user commands. The facial expressions are captured by means of a digital camera and interpreted by an application running on a laptop computer on the wheelchair. The software includes digital image processing algorithms for feature detection, such as colour segmentation and edge detection, followed by the application of a neural network that uses these features to detect the desired facial expressions. A specific command language for controlling the intelligent wheelchair was also developed in the scope of this work, as well as a simple wheelchair simulator for validating the final results.
