Real-time robotic avatar control using fuzzy gaze-classification for people with disability

Abstract

Robotic control with gaze-classification has been an area of interest for quite some time. In this paper, a novel solution of implementing sensing system with intelligent visualization has been presented. Such a system can have a variety of applications but can be of exceptional importance for patients suffering with spinal muscular atrophy (SMA) disorder, which gradually end with paralyzed state. The presented solution incorporates Optical Gaze System, within a Virtual reality (VR) headset to control a robotic Avatar remotely based on the gaze position. The robot is being called as avatar due to its ability to maneuver and provide audio and video streams to the user wearing the VR headset, thus acting as an extension of the user him/herself. Hence people with permanent disability due to degenerative diseases such as SMA can benefit from this design by utilizing the avatar to see things and places where they cannot go themselves. The psychological impetus can assist tremendously in their treatment. The Fuzzy Algorithm presented here enables a very fast image classification method for such near-real-time application. Consequently, the processing and decision making of the system takes about a second with an accuracy of over 95%. 2015 IEEE.