Detection of Driver Fatigue Based on Multi-physiological Signals in Wireless Body Area Network

FU Rong-rong,WANG Hong,WANG Lin,ZHANG Chi

DOI: https://doi.org/10.3969/j.issn.1005-3026.2014.06.020

2014-01-01

Abstract:Driver fatigue was detected using electroencephalograph, electromyography and respiration signals,w hich w ere collected w irelessly. The approximate entropies of the three signals w ere selected as features,and the reduction of feature dimensions w as achieved by principle component analysis. Statistical analyses w ere then given to both original features and principle components and an evaluation model for driver fatigue w as established using regression equation. The experimental results w ere evaluated by cross validation and the accuracy w as more than 90% based on data fusion method. The results verify that the model is effective in detecting driver fatigue.