Fatigue-driving recognition based on the state of the human eye and electroencephalographic signal

The paper apply dual indicators based on eye's state and electroencephalographic signals to identify whether the driver is at the state of fatigue or not. This method improves the detection rate and detection precision, reduce error detection rate, and increase system's robustness and stability compared with single eye movement signal or electroencephalographic signals. The first step is to collect eye's state information, then use the PERCLOS algorithm to identify whether a driver is at the state of fatigue or not. At the same time the power ratio of alpha, beta and theta which are received by parallel processing EEG signal identifies whether the driver is at the state of fatigue or not. As long as the system meets one of the two indicators, the driver is identified to fatigue and the system make alarm.