基于脑电频域特征的高速铁路调度员监控工作注意水平识别方法  

Attention Level Recognition of High-speed Railway Dispatchers’Monitoring Work Based on EEG Frequency Domain Features

作　　者：张光远[1,2,3] 文原劲 李婧 章子睿[1,2,3] 胡悦 ZHANG Guangyuan;WEN Yuanjin;LI Jing;ZHANG Zirui;HU Yue(School of Transportation and Logistics,Southwest Jiaotong University,Chengdu 610031,China;National United Engineering Laboratory of Integrated and Intelligent Transportation,Southwest Jiaotong University,Chengdu 610031,China;National Engineering Laboratory of Integrated Transportation Big Data Application Technology,Southwest Jiaotong University,Chengdu 610031,China)

机构地区：[1]西南交通大学交通运输与物流学院,四川成都610031 [2]西南交通大学综合交通运输智能化国家地方联合工程实验室,四川成都610031 [3]西南交通大学综合交通大数据应用技术国家工程实验室,四川成都610031

出　　处：《铁道学报》2023年第5期21-28,共8页Journal of the China Railway Society

基　　金：四川省科技计划项目(2021YJ0043);西南交通大学本科教育教学研究与改革项目(20221103,2103065)。

摘　　要：高速铁路调度员监控作业的注意力水平识别是全周期注意识别的重要组成部分。针对监控作业交互少、反馈弱以及视觉特征不显著的特点,设计基于信息感知密度的注意诱导实验作为客观评价参照,采集全头脑电并提取了57个通道的7项频段指标为识别特征。采用Pearson相关系数进行特征初筛,采用Logistic回归-预测变量重要性排序的包裹式方法对特征进行进一步降维并进行注意水平识别。实验结果表明:基于左额叶和双侧枕叶的17个脑电频段特征的多项Logistic回归模型对低注意水平有81%的识别准确率。脑电频段特征对应负责大脑思维功能和视觉加工处理的脑功能区,反映高铁调度员在监控工作中的注意水平变化对应的认知功能变化。The attention level recognition of high-speed rail dispatchers monitoring operation is a vital part of the full-cycle attention recognition.To address the challenges of less interaction,weak feedback,and insignificant visual characteristics of monitoring tasks,an attention induction experiment based on information perception density was designed as an objective evaluation reference.The whole EEG was collected and 7 frequency band indicators from 57 channels were extracted as identification features.The Pearson correlation coefficient was used for feature preliminary screening,and the package method of Logistic regression-predictor importance ranking was used to further reduce the dimensionality of features and identify the level of attention.The experimental results show that the multinomial Logistic regression model based on 17 EEG features of the left frontal lobe and bilateral occipital lobe has an accuracy of 81%for the identification of low attention levels.The EEG frequency band characteristics correspond to the brain functional areas responsible for thinking function and visual processing,and reflect the changes in cognitive function corresponding to the changes in attention level of high-speed rail dispatchers in monitoring work.

关 键 词：高速铁路 调度员 监控工作 脑电 注意力水平 多项Logistic回归 

分 类 号：U292.4[交通运输工程—交通运输规划与管理]