Fatigue Driving Detection with Artificial Bee Colony Algorithm

Fatigue driving detection (FDD) with electroencephalogram (EEG) is an effective manner for ensuring the driving safety. As EEG signals used for recognition contain a large amount of information, they are always time-consuming to be processed. In order to improve the efficiency and accuracy of FDD, 31 features are extracted from time domain, frequency domain, time-frequency domain respectively in this paper, and a dimension reduction strategy (FDR) is introduced in the initialization stage of the artificial bee colony (ABC) algorithm to reduce the number of features. The results show that the efficiency of the proposed algorithm is 32.5% higher than that of traditional one for mental fatigue detection.