基于实车试验的驾驶人换道行为多参数预测

发布时间：2018-03-24 12:20

本文选题：交通工程　切入点：辅助安全系统　出处：《长安大学学报(自然科学版)》2014年05期

【摘要】：针对驾驶人在换道时若出现决策失误,极易引发交通事故的问题,通过在真实交通环境中进行实车试验,采集车辆运动状态、驾驶人操作行为以及头部运动特性、周围交通环境等数据;通过对换道意图阶段和车道保持阶段数据参数的对比分析,提取能够表征驾驶人换道意图和行为的特征参数;通过建立BP神经网络模型,以不同特征参数作为输入向量对待测样本进行预测,确定最终的输入特征指标,并基于建立的BP网络模型,进行驾驶人换道行为预测。研究结果表明:换道前2s内的预测准确率为94.4%,灵敏度为93.33%,能够准确预测出93.33%的换道行为;该模型能够有效预测驾驶人的换道行为,且准确率高、时序性强。
[Abstract]:In view of the problem that the driver is prone to cause traffic accident if he makes a wrong decision when changing the road, the real vehicle test is carried out in the real traffic environment to collect the moving state of the vehicle, the operating behavior of the driver and the characteristics of the head motion. Traffic environment and other data; through the comparison and analysis of the data parameters in the phase of road change intention and lane maintenance, the characteristic parameters which can represent the change intention and behavior of the driver are extracted; the BP neural network model is established, and the BP neural network model is established. Different characteristic parameters are used as input vectors to predict the test sample, and the final input feature index is determined, and based on the BP neural network model, The results show that the prediction accuracy is 94.4 and the sensitivity is 93.330.The model can predict 93.33% of the changing behavior, and the model can effectively predict the changing behavior of the driver, and the accuracy is high. Strong timing.
【作者单位】：长安大学汽车学院;长安大学汽车运输安全保障技术交通行业重点实验室;
【基金】：教育部长江学者与创新团队支持计划项目(IRT1286) 国家自然科学基金项目(61374196,51178053) 中央高校基本科研业务费专项资金项目(CHD2012TD006,2013G1221028)
【分类号】：U491.25

【参考文献】