期刊论文

Fatigue Driving Detection Based on Autoregressive Conditional Duration Model

Mao, S.-H.

中文

交通运输系统工程与信息

1009-6744

2018

18

3

81-87

10.16097/j.cnki.1009-6744.2018.03.013

国家自然科学基金

本校为第一且通讯机构

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 sp...