摘要
结合N-back任务范式和持续操作测试(CPT)任务范式,设计出模拟认知性VDT持续监控作业的实验.根据实验数据:分别对脑力负荷评估指标体系中的6个评估指标在20个作业时间段的均值进行方差分析,差异显著;训练非线性自回归神经网络模型(NAR),对不同作业时间段脑力负荷在评估指标上发生的变化进行动态预测;再结合认知性VDT持续监控作业人因可靠性评估模型对人因可靠性概率进行预测.研究结果表明:该方法可动态预测不同时间段作业者的人因可靠性,实现认知性VDT持续监控作业任务的动态分配,提高系统可靠性.
The experiment was designed base on N-back paradigm and CPT paradigm, which can simulate cognitive monitoring tasks. By using variance analysis method, we found that the mean values of the 6 mental workload assessment indicators were significantly different in the 20 working time periods. 6 nonlinear auto regressive(NAR) networks were trained for dynamic prediction of 6 indicators’ value, and then the possibility of human reliability could be predicted by combining the human reliability evaluation model. The result shows that the method can predict the human reliability and assign task of VDT cognitive monitoring operating systems dynamically.
作者
冯海芹
廖斌
罗俊浩
王泰鑫
FENG Hai-qin;LIAO Bin;LUO Jun-hao;WANG Tai-xin(Automotive and Information Engineering Department,Urban Vocational College of Sichuan,Chengdu 610101,China;School of Business,Sichuan Normal University,Chengdu 610101,China)
出处
《数学的实践与认识》
北大核心
2019年第4期99-105,共7页
Mathematics in Practice and Theory
基金
四川省教育厅科研项目(18ZB0346)
教育部人文社会科学研究青年基金(14YJCZH089)
四川省社科规划项目(SC16B065)
四川师范大学开放实验项目(KFSY2018038)
关键词
VDT
认知性监控
人因可靠性
动态预测
非线性自回归神经网络
visual display terminal(VDT)
cognitive monitoring
human reliability
dynamic prediction
nonlinear auto regressive(NAR)
