EEG-Based Driver Distraction Detection via Game-Theoretic-Based Channel Selection

In recent years, there has been much effort to estimate drivers’ state with the goal of improving their driving behavior and preventing vehicle crashes in the first place. Physiological based detection has shown to be the most direct method of measuring driver state among which, electroencephalogram (EEG) is the most comprehensive method. However, EEG-based driver state detection faces the challenge of computational complexity of data mining algorithms given high density and resolution of EEG signals recorded from multiple channels. On the other hand, in order to early detection and prevention of driver critical states real-time responsiveness of the monitoring system is necessary. This challenges can be tackled by localizing the regional impact by selecting a small subset of coherent channels and reducing the processing load on all channels. In this paper, we present and investigate a Game-Theoretic-Based approach for EEG channel selection, in order to localize the most efficient sub-set of channels in addition to maximizing the driver distraction detection accuracy. In this way, we apply game theory based channel selection algorithm based on the utility measure, Shapley value, in exact to estimate overall usefulness of each EEG channel. We then consider the combination of channels and evaluate their performance. Empirical comparison of best combination of channels, best ordered channel based on Shapley value with another existing feature selection method shows that the sub-set of channels leads to the best detection performance in terms of accuracy (90.12% accuracy).