Development of a wearable eye tracker to control an autonomous wheelchair

Chan, Huaxuan (2021) Development of a wearable eye tracker to control an autonomous wheelchair. Final Year Project, UTAR.

Preview

PDF Download (1339Kb) | Preview

Abstract

In this study, a low-cost wearable eye-tracking system was developed to capture gaze information of the user and used to navigate an autonomous wheelchair. The eye-tracking system utilizes the video oculography method to record movements of the eye and to determine points of interest of the user. The development of this Human-computer-interface between the eye-tracker and the wheelchair would help people with severe motor disabilities to control the wheelchair and regain certain degrees of autonomy in locomotion. Existing research on low-cost wearable eye trackers lacks attention on the product design, which contributes to slippage effects and causes discomfort for the user. This study aims to redesign and optimize an existing eye-tracker developed the previous year by another student for greater cost efficiency, higher robustness, and to achieve deviation of angular accuracy within 1°. The prototype would be evaluated using the Pupil Capture software where parameters of accuracy, precision and data loss can be obtained. The developed eye-tracker is lighter and cheaper (69.5 gms and RM132.00) compared to an earlier model (138.4 gms and RM330.00). The deviation in angular accuracy and angular precision of the developed eye-tracker system was within the 1.5° to 2.5° range that was slightly higher than that of commercially available eye-trackers of within 1o which costs more than RM10,000.00 each. However, further developments of the prototype are required for its intended application of controlling an autonomous wheelchair.

Actions (login required)