基于EEG的驾驶持续性注意水平PSO-SVM识别模型

为了对驾驶持续性注意水平予以有效识别,基于脑电( EEG)信号特征指标构建了一种持续性注意水平识别方法. 以驾驶行为绩效为客观测评指标,提出了一种驾驶持续性注意水平等级划分方法. 在此基础上,选取驾驶员EEG波段 (θ(4~8 Hz)、α(8~13 Hz)、β(13~30 Hz))的频谱幅值及其组合指标(α+β)/β、α/β、(θ+α)/(α+β)、θ/β、(α+β)/θ作为特征指标,将粒子群优化( PSO)算法与支持向量机( SVM)相结合,构建了驾驶持续性注意水平识别算法. 最后,基于驾驶模拟器实验数据对该模型予以试算. 结果表明模型识别平均正确率可达93. 02℅.该方法可用于对驾驶员持续性注意水平的识别.

PSO-SVM Identification Model for Driving Sustained Attention Level Based on EEG

In order to recognize driving sustained attention effectively, an identification method for sustained attention level was proposed based on the signal of electroencephalograph ( EEG ) . Firstly, taking the driver' s reaction time to random events as indexes, a dividing method for sustained attention levels was proposed. Secondly, using average spectrum amplitude from the bands of (θ(4~8 Hz) ,α(8~13 Hz) ,β(13~30 Hz) ) of EEG and its' ration value (α+β)/β,α/β, (θ+α)/(α+β) ,θ/βand (α+β)/θ as characteristic indexes, combining the particle swarm optimization ( PSO ) with support vector machine (SVM),an identification model for identifying sustained attention level was proposed. Finally, based on the data from driving simulating, the identification model was tested. The result shows that the average accuracy rate of model is 93. 02℅ and the method is applicable to identification of driving sustained attention level.