EEG-based fatigue recognition using deep learning techniques

• Main Author: Chua, Zhong Sheng
• Other Authors: Alex Chichung Kot
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2022
• Subjects: Engineering::Electrical and electronic engineering
• Online Access: https://hdl.handle.net/10356/157409
• Institution: Nanyang Technological University

Fatigued driving has always been a factor for traffic accidents, and it has prompted an interest in detecting driver’s fatigue. A variety of methods has been proposed and Electroencephalogram (EEG)-based mental state analysis is a reliable and effective way to detect fatigue. With the advancement of Deep machine learning, it has gained attention for producing better results than the standard approach. This paper proposes using a feature extraction which uses Autoregression (AR) to extract characteristics of the EEG signals and then process to into a classification algorithm which Convolution Neural Network (CNN) would be used. The results from another published paper using the same dataset will be utilized as a baseline for performance comparison. The proposed method would use a single channel baseline comparison and a leave one subject out validation to ensure that the actions performed are same. In comparison to the baseline, our proposed method has a mean classification accuracy for detecting fatigue at 69.59 %.