应用于换道预警系统的潜在危险目标辨识算法

Potential dangerous target identification algorithm for lane change warning system

针对换道预警系统中的潜在危险目标辨识问题,搭建了车辆运动状态监测试验车,获取了多名被试驾驶人在实际道路上的自然驾驶数据。采用车速与车身横摆角速度建立了道路曲率估算模型,选取车道线作为模型的参考对象,提出了基于车道线的车辆之间位置关系辨识模型,来确认换道预警系统的潜在危险目标。研究结果表明:曲率估算模型误差较大,计算曲率极值的相对误差达到了10.2%,但车辆之间车道关系辨识模型能够准确辨识出潜在危险目标,对其他车辆换道过程的横向越线时间计算精度优于0.1s;当曲率估算相对误差达到28.4%时,相对车道关系辨识算法依然能够准确识别出其他车辆所处的车道信息。

Aiming at the potential dangerous target identification problem of lane change warning system,motion state test vehicle was set up,and real road driving data of the test vehicle and other vehicles were obtained.Speed and yaw rate were used to establish road curvature-estimating model.Based on this model,according with lane marks,lane related vehicles position identification algorithm were proposed in order to distinguish the potential dangerous target for lane change warning system.Calculation result shows that the error of road curvature estimating model is larger,and the extremum of relative error achieves 10.2%.But the relative lane identification model can recognize the potential dangerous targets accurately.Error of Time calculated by this model for other vehicles line-crossing during lane change process is better than 0.1second.When the relative estimating error of road curvature achieves 28.4%,the relative lane position distinguish model can still identify the lane information of other vehicles accurately.