应用于换道预警的驾驶风格分类方法

针对不同驾驶风格驾驶人对换道预警需求存在差异的问题，采用视觉传感器、 雷达传感器、GPS、车辆 CAN 总线数据采集系统，基于小型乘用车搭建了实际道路驾驶行 为试验车.通过对多名驾驶人进行实际道路自然驾驶试验，选用跟车时距、换道时距、超速 频次、换道过程最大方向盘转角等参数，采用模糊评价法建立了驾驶风格离线分类模型， 实现了将驾驶人分类为冒进型、比较冒进型、比较谨慎型、谨慎型四类.采用换道安全性相 关参数对模型进行验证. 结果表明，随着驾驶人谨慎程度的增加，换道安全性评价参数更 倾向于安全，驾驶风格分类模型与实际换道安全性评价参数之间呈现良好的一致性.

Driving Style Classification Method for Lane Change Warning System

Aiming at the different requirement of lane change warning system for different drivers with dif- ferent driving style, vision sensor, radar, GPS, vehicle CAN bus data capture system are installed in a small passenger car, and real road driving test is carried out. Based on this, real road driving data of different driv- ers is collected. Following headway, lane change headway, overspeed frequency and max turning angle of steering wheel during lane change are selected as judge parameters, and fuzzy evaluation method is used to establish classification model of driving style in the way of offline. Drivers are divided into aggressive driv- er, relatively aggressive, relatively cautious and cautious. Parameters related to lane change safety are used to verify the classification model, and the results show that with the increasing of cautious level, the lane change safety parameters are inclined to safely correspondingly. The consistency between driving style classi- fication model and lane change safety real judge parameters is good.