摘要
为探寻操作手机打车软件时的驾驶分心识别方法,本文开展模拟驾驶实验,采集了驾驶员在不同驾驶状态下的驾驶行为参数,通过对参数的统计分析,确立分心检测参数集。采用支持向量机分类算法理论构建基于驾驶绩效的分心检测模型,并利用实验数据验证模型的有效性。结果表明:该模型对驾驶员视觉分心驾驶行为检测率最高,正常驾驶行为次之,对认知分心驾驶行为的检测能力最弱,模型的平均检测正确率为86.67%,检测效果较好,可用于驾驶分心检测。
For efficient detection of distracted-driving when using smartphone taxi-hailing applications, the paper carries out simulated driving experiments and collects the driving behavior parameters under different driving conditions. The set of distracting parameters is established via a statistical analysis of the collected parameters. Support vector machine is then used to construct the distracted-driving detection model based on the driving performance. The proposed model is verified using the experiment data. The results show that the model has the highest detection rate for visual-related distracted driving, followed by normal driving behavior, and cognitive-related distracted driving. The average detection rate of the model is 86.67%.
出处
《交通运输工程与信息学报》
2018年第1期9-14,31,共7页
Journal of Transportation Engineering and Information
关键词
驾驶分心
驾驶绩效
支持向量机
检测模型
distracted-driving
support vector machine
driving performance
detection model
