Publication

Abstract :

Various research studies conducted all around the world reveal that 75% of people with debilitating physical disabilities and elderly people experience significant difficulties for their day to day locomotion. Although there are existing solutions to overcome this problem, none of them served to be customizable, economical and user friendly all at the same time. Therefore, one of the best options that could be useful for these affected class of people is a gesture-based interaction with their environment, in particular their wheelchairs. The proposed system named as 'Gest-BOT', uses a small camera mounted very close to the user's hand, which tracks the small movements of their fingers to understand the direction of movement of the wheelchair. A gesture recognition system which identifies the gesture is then interfaced to the wheelchair control system in order move it to the desired location. Gest-BOT is effortless to use, customizable, economical, highly convenient and non-intrusive. The hardware implementation of Gest-BOT consists of a wheelchair, an Arduino board based ATmega microcontroller and an Intel Atom EBC-352 processor, which uses MATLAB® executables for gesture image processing. We have also calculated the response of the system under various conditions which is detailed in the paper. © 2012 IEEE.