基于EEG熵值的驾驶员脑力负荷水平识别方法

Recognition method of driving mental workload based on EEG entropy

为了对驾驶员脑力负荷予以有效识别,基于脑电信号指标构建了一种驾驶员脑力负荷识别方法.对驾驶员脑电信号进行快速傅里叶变换（FFT）,选取θ（4~8 Hz）,α（8~13 Hz）,β（13~30 Hz）3个频段的频谱幅值分别进行熵处理,对所得到的熵值作为脑力负荷识别参数,并对识别参数进行Kruskal-Wallis检验,选取差异最为显著的10项参数作为脑力负荷特征指标,在此基础上结合BP模型构建了驾驶员脑力负荷识别模型.基于驾驶模拟器实验数据,模型识别正确率为87.8%~90.4%.结果表明,该模型对驾驶员脑力负荷识别具有较高准确性,可实现不同驾驶员脑力负荷的有效识别,为未来自动辅助驾驶系统构建及车载信息系统优化设计提供算法依据.

In order to recognize driving mental workload efficiently,a recognition method of driving mental workload based on EEG indices is constructed.After the fast Fourier transform （FFT）of the electroencephalograph （EEG）,the entropy processing of three bands of spectrum,θ（4 to 8 Hz）,α（8 to 13 Hz）,β（13 to 30 Hz）,are conducted respectively,and the value of entropy is used as mental workload recognition parameter.Then 10 difference-remarkable indices are chosen as the characteristic features after the Kruskal-Wallis test of the recognition parameters.Meanwhile,combi-ning with the back propagation （BP）neural network,the recognition model for state of driving mental workload is established.The EEG data based on the simulator are used to test the model and the recognition accuracy rate is within 87.8% to 90.4%.The results show that the proposed model is accurate for the recognition of driving mental workload and achieves the recognition of different drivers.The model provides a algorithm basis for constructing automatic auxiliary driving system in the future and the optimization design of the traffic information system.