Research On Driving Fatigue Detection Based On EEG Signal Recognition

Now,fatigue driving has become the fifth leading cause of road safety hazards.Fatigue driving detection based on physiological electrical signals has become a hot topic in current research.Brain-computer interface(BCI)is a technology that controls external devices only through brain potential activity not by relying on peripheral nerves.With the popularization of BCI technology and the development of signal processing technology,fatigue driving detection based on EEG signals has become possible.The main research work of this paper is as follows:(1)In order to fully extract the feature of fatigue driving EEG signals,a feature extraction method combining the power spectral density(PSD)algorithm and the ensemble empirical mode decomposition(EEMD)method is proposed.Recruit 6 healthy subjects for simulated fatigue driving experiments and acquire 32-lead EEG signals.After the preprocessed,EEMD decomposition is performed to obtain several Intrinsic Mode Functions(IMF),and the powers of the IMF components obtained by PSD feature extraction.The experiment results show that,the new method is more suitable for the feature extraction of the fatigue driving EEG signal,which makes the classification performance have a qualitative improvement.(2)When hierarchical extreme learning machine(H-ELM)works,the hidden layer penalty factor C and the number of hidden layer K are arbitrary,which make it can not adjust the model adaptive under different data conditions.An artificial fish-swarm algorithm(AFSA)optimized H-ELM classifier is proposed.The experimental results show that the classification performance of the AFSA-H-ELM is better than the traditional classification,which provides a new idea for the identification of fatigue driving EEG signals.(3)The proposed methods are compared experimentally to prove the superiority of the method.Firstly,compared the AFSA-H-ELM algorithm with the traditional classification algorithm.The results show that the classification performance of AFSA-H-ELM is better than original algorithm.Secondly,compared two different feature extraction algorithms and send feature information to different classifiers,including the AFSA-H-ELM.The results demonstrate that the proposed feature extraction algorithm is superior to the traditional PSD algorithm when it is used to classify the EEG signals of fatigue driving,and the better recognition accuracy is achieved when classifying EEG signals by AFSA-H-ELM.