Real Time Drowsiness Detection Based on Lateral Distance Using Wavelet Transform and Neural Network

The paper proposed a model using real time driving front video recording to detect driver drowsiness. The video recordings were fed into the TRW's simulator to obtain the lane-related signals. Time domain features and frequency domain features were extracted from the lane-related signals to characterize the difference of alert state and drowsiness state. Both support vector machine and neural network were used to detect the drowsiness. Experimental results on real word driving recordings illustrated that the proposed method can detect the drowsiness with good accuracy. It also show that TRW simulator can generate reliable lane related signals if high quality video sequences are provided.