VEHICLE DATA ACQUISITION AND ANALYSIS FOR DETECTION OF THE DRIVER'S DROWSINESS

In the last decade a big effort was made to develop a reliable system for detection of the oncoming driver's sleepiness. Until now this problem has not been satisfactorily solved. In this paper we present our system for data acquisition and analysis. The data is collected from the engine CAN bus during test-rides in the real traffic conditions. The driver drowsiness is measured directly in the vehicle before and after the test-ride using the fusion of two methods: flicker test and subjective feeling. The data is analyzed using the data mining techniques including feature extraction and filter feature selection. The features are based on the discrete wavelet transform and short time Fourier transform. The performance of the features is measured by the area under the receiver operating characteristic.