摘要
针对危险驾驶人群进行了实时驾驶风格辨识方法研究,提出了考虑驾驶人与周围环境交互的驾驶风格识别模型。以模拟驾驶采集数据为基础,以3 s为时间窗口,将实时横向偏移量、跟车距离、驾驶人注视点位置转化为驾驶风格表征因子,运用FCM-M(Fuzzy C-mean-M)对驾驶人驾驶风格进行聚类,将实时驾驶风格标定为冲动型、较冲动型、普通型、保守型,并构建神经网络模型对实时驾驶风格进行识别。结果表明:驾驶人驾驶风格波动与驾驶人注视点转变有密切联系,神经网络模型对驾驶风格识别准确率最高可达到99.1%,表现出良好的预测效果。
A real-time driving style identification method for dangerous drivers is studied.A driving style identification model considering the interaction between drivers and the surrounding environment is proposed.Based on the data collected by simulated driving,3 s as the time window.Real-time lateral offset,following distance and driver’s gaze point position are transformed into driving style characterization factors.FCM-M(Fuzzy C-mean-M)is used to cluster drivers’driving styles.The real-time driving style is classified as grumpy,impulsive,ordinary and conservative.Construct neural network model to identify real-time driving style.The results show that the fluctuation of driver’s driving style is closely related to the change of driver’s gaze point.The neural network model has the highest recognition accuracy of 99.1%for driving style,which shows a good prediction effect.
作者
葛慧敏
黄嘉慧
臧文凯
董磊
周礼军
GE Huimin;HUANG Jiahui;ZANG Wenkai;DONG Lei;ZHOU Lijun(School of Automotive and Transportation Engineering,Jiangsu University,Zhenjiang 212013,China)
出处
《重庆理工大学学报（自然科学）》
北大核心
2023年第8期177-184,共8页
Journal of Chongqing University of Technology：Natural Science
基金
国家自然科学基金项目(51905224)。
关键词
交通工程
交互
驾驶风格
模糊聚类
识别模型
traffic engineering
interaction
driving style
fuzzy clustering
identification model
