车辆跟驰情景下酒后驾驶行为动态特征参数提取  

Dynamic Characteristic Parameters Extraction of the Drunk Driving in the Vehicle Following Sence

作　　者：孙一帆[1] 张敬磊[1] 王丝丝[1] 

机构地区：[1]山东理工大学交通与车辆工程学院,淄博255000

出　　处：《科学技术与工程》2017年第28期144-149,共6页Science Technology and Engineering

基　　金：国家自然科学基金(61573009);山东省自然科学基金(ZR2017LF015);山东省高等学校科技计划(J15LB07);汽车安全与节能国家重点实验室开放基金(KF16232)资助

摘　　要：酒驾是导致重大交通事故的原因之一,酒驾辨识已成为交通安全研究中的重要问题。因此,提取不同驾驶倾向性驾驶员在跟驰状态下的酒驾特征参数对实现酒驾的准确辨识有重要意义。由问卷调查确定驾驶员驾驶倾向性;通过人因工程实验、驾驶模拟实验,采集不同类型驾驶员的生理、操作行为、车辆运行、驾驶环境等动态信息;采用神经网络分类器,获得备选特征集合的分类正确率估计,运用离散粒子群算法提取不同驾驶倾向性驾驶员的酒驾特征参数。验证结果表明提取的特征参数能有效辨识不同驾驶倾向性驾驶员是否酒驾。Drunk drying is a reason of serous traffic accidents, the identification of drunk drying has becomean important issue of traffic safety study. It’s very important for realizing the accurate idening to extract drunk driving characteristics of different driving propensity drivers in the vehicle following scene. D e -termining the dryers propensity by questionnaire s u r e y; to gather different type drivers’ dynamic information ofdriver physiology, operation behavior, vehicle m o v e m e n t,driving conditions by human factors engineering experi-ment and driving simulation experiment; the classification accuracy estimation of optional tained by neural network, discrete particle swarm optimization algorithm was used to extract drunk driving charac-teristic parameters of different driving propensity drivers. Verification results showthat the extracted characteristicparameters can identify whether the different driving propensity driver is drunk driving or not.

关 键 词：交通安全 酒后驾驶行为 驾驶倾向性 特征提取 

分 类 号：U491[交通运输工程—交通运输规划与管理]