Driver fatigue Assessment Based on the Feature Fusion and Transfer Learning of EEG and EMG

The methods for driving fatigue assessment based on multiple physiological signals have received extensive attention in recent years. In view of the limitations of individual differences leading to reduce the accuracy of fatigue assessment, this paper proposed a method for driving fatigue assessment based on the feature fusion and transfer learning of electroencephalogram(EEG) and electromyography(EMG). The EEG and cervical EMG signals from 20 healthy subjects were recruited synchronously using the simulated driving platform. The time-frequency analysis method was used to extract the fatigue characteristics of EEG signals and EMG signals. A driver fatigue assessment model based on feature fusion and transfer learning was constructed to achieve effective identification of different drivers' fatigue state. The results showed that the recognition rate of driving fatigue based on the feature fusion and transfer learning of EEG and EMG was higher than traditional support vector machine method, and it could improve the robustness of driving fatigue assessment in different subjects. Therefore, this paper provides a new method for practical application of driving fatigue assessment based on physiological signal.