基于自回归条件持续期模型的疲劳驾驶研究  

Fatigue Driving Detection Based on Autoregressive Conditional Duration Model

作　　者：毛树华[1] 王先朋 文江辉[1,2] 吴超仲[2] 肖新平[1] MAO Shu-hua;WANG Xian-peng;WEN Jiang-hui;WU Chao-zhong;XIAO Xin-ping(a.College of Scienc;b.Intelligent Transport Systems Research Center,Wuhan University of Technology,Wuhan 430070,Chin)

机构地区：[1]武汉理工大学理学院,武汉430070 [2]武汉理工大学智能交通系统研究中心,武汉430070

出　　处：《交通运输系统工程与信息》2018年第3期81-87,107,共8页Journal of Transportation Systems Engineering and Information Technology

基　　金：国家自然科学基金(51479151;61403288)~~

摘　　要：对实际驾驶实验中不同驾驶员的车速数据进行处理,得到车速变化持续期间序列,应用自回归条件持续期模型(ACD),讨论了车速变化持续期的相关性质,并对模型的可靠性做了评估.使用ACD模型为车速变化持续期时间序列建模,其优点是能够在不损失原始非等间隔时间序列特性的条件下,直接分析得到驾驶状态的时域微观性质.采用EACD(1,1)和WACD(1,1)模型对不同驾驶员的车速变化时间序列进行建模,结果表明:其具有较好的拟合程度,当实际期间小于条件预期期间,驾驶员的驾驶状态有变好的可能;当实际期间大于条件预期期间,驾驶员的驾驶状态有变差的可能.For driving speed data of different drivers in real driving experiments, this paper gets the sequence of vehicle speed change duration, applies Autoregressive Conditional Duration Model（ACD） to discuss the related properties of vehicle speed change duration, and evaluates the reliability of the model. The ACD model is used to model the time series of vehicle speed change duration. The advantage is that it can directly study the microscopic properties of driving states without losing the characteristics of original non equal interval time series. Using EACD（1,1） and WACD（1,1） to model the speed change time series of different drivers, the results show that it has good fitting effect. When the actual duration is smaller than the expectation of conditional duration, the driver＇s driving state may become better, and when the actual duration is more than the expected duration, the driving state is likely to be bad.

关 键 词：交通工程 疲劳驾驶 ACD模型 车速变化持续期 准极大似然估计 Nelder-mead单纯形法 

分 类 号：U491.254[交通运输工程—交通运输规划与管理]