Fuzzy Fusion of Eyelid Activity Indicators for Hypovigilance-Related Accident Prediction

Abstract

In this paper, a fuzzy expert system (FES) for the detection of the physiological manifestations of extreme hypovigilance is presented. A large number of features that describe the eyelid activity of drivers is examined, and fuzzy logic is used for the fusion of the most prominent features to not only increase the accident prediction accuracy but also provide a reliable system that generates a small number of false warnings. For the development and testing of the system, driving simulator data from 35 drowsy subjects were used. In addition, a secondary control group of 13 alert drivers was used for the estimation of the trained system's false alarm ratio. The results show that a fuzzy combination of eyelid activity parameters may lead to a system with high sensitivity and specificity in predicting sleep onset and related accidents.