SelfBOT: An Automated Wheel-Chair Control Using Facial Gestures Only

Facial gestures serve as a promising means of non-verbal communication, as effectively demonstrated in applications like cursor movement, eye tracking, and eye interaction. Our study introduces a novel facial-interaction-based wheelchair system designed to enhance the mobility and independence of physically disabled individuals who rely on assistance for navigation. As a proof-of-concept, this application detects facial, nose, and eye movements to monitor user actions and execute corresponding operations. A novel algorithm accurately detects facial movements up, down, left, right, blinks, and eye flickers, enabling users to control the wheelchair’s movement in any direction, including clockwise and counterclockwise. Real-time feedback indicates whether facial gestures successfully control wheelchair movement. Our system maintains accurate gesture recognition under varying environmental conditions and provides smooth, intuitive cursor control, a user-friendly interface, and a cost-effective, energy-efficient implementation that sets our system superior apart from existing solutions. It achieves a detection accuracy of approximately 96% in well-lit environments, surpassing existing systems. While accuracy drops slightly to 83% in dark environments, it still exceeds existing capabilities. This system empowers physically disabled individuals, enabling those with quadriplegia, amyotrophic lateral sclerosis(ALS), cerebral palsy, or spinal cord injuries to control their wheelchairs using facial gestures characterized by its unique ability to move in any direction as opposed to just four or eight directions.