基于脑电的乘员信息处理作业脑力负荷状态识别模型

摘要

脑力负荷状态的准确识别是装甲车辆乘员信息处理作业行为研究的关键技术,对提高人机系统的作战效能具有重要意义.针对乘员作业类型向信息处理作业转变的基本趋势,提出了融合小波包分解(WPD)和快速独立分量分析(FastICA)的脑电信号预处理方法,建立了反映脑力负荷状态的EEG信号特征空间.基于粒子群优化(PSO)算法和支持向量机(SVM)构建了乘员信息处理作业脑力负荷状态识别模型;并面向目标录入典型信息处理作业对识别模型进行了实例应用,旨在为解决乘员信息处理作业脑力负荷的准确识别探索新的途径.结果表明,该模型脑力负荷状态识别的平均正确率可达96%,可实现不同乘员脑力负荷的量化识别,具有良好的预测精度和可重用性.%The veracity of operation mental workload recognition is essential to the study of operation behavior in armored vehicle cabin crew's information processing, which has important implications for improving operational effectiveness of the man-machine system.In view of the basic trend that the task type gradually changes to information processing task,the artifact removal method based on fast independent component analysis combined with wavelet packet decomposition (FastICA-WPD) was proposed,the feature space was built, the recognition model of mental workload based on particle swarm optimization combined with support vector machine (PSO-SVM) was built, the model was verified by an example, with an aim to solve the problem of crew's information processing mental workload recognition.The results indicate that the recognition accuracy rate is 96%,could achieves the recognition of different crews,the recognition accurately and recognition reusability are preferably.