基于脑电熵值特征和功能连接的不同线型道路下驾驶状态检测OACSCDCSTPCD

The decoding of different driving processes and driving intention prediction based on electroencephalo-gram(EEG)signals are the key issue of human computer interface based intelligent driving control.In order to realize the identification of driving process in straight road,left curve and right curve road,this paper proposes a feature extrac-tion method for driving behaviors based on functional brain network and entropy features of EEGs,and achieve thes clas-sification of different driving conditions under various curve by combining with several classifiers with the extracted EEG features.Corresponding simulation driving experiments are designed and the results show that the method proposed in this study can effectively identiy the driving process under various curve.All the classification accuracy for the 16 subjects between straight and curve road are higher than 82％,with the highest value to 86.66％,while the accuracy between left curve and right curve are higher than 75％,with the highest value to 77.95％.Interdependence analysis between different brain regions shows that there are obvious brain contralateral characteristics during curve driving.In addition,left curve driving needs more interactions between the brain regions than right curve driving,while the left hemisphere is slightly more active than the right hemisphere during straight driving.The result has significant value for understanding driver's brain cognitive characteristics in the process of curve driving and driving behavior detection under different curve road.