基于Logistic模型的驾驶人换道意图识别方法.pdfVIP

科技导报2014，32（14） 基于基于LogisticLogistic模型的驾驶人换道意图模型的驾驶人换道意图 识别方法识别方法 1 2 1 1 彭金栓 ，付锐 ，邵毅明 ，徐磊 1.重庆交通大学山地城市交通系统与安全重庆市重点实验室，重庆400074 2. 长安大学汽车运输安全保障技术交通行业重点实验室，西安710064 摘要摘要 为有效降低车道变换行为诱发事故的风险性，提出一种基于Logistic模型的驾驶人换道意图识别方法。利用faceLAB视 觉追踪系统，通过真实环境下的实车测试，结合换道前驾驶人对后视镜的注视特性确定换道意图时窗，分析车道保持与换道意图 阶段的注视特性差异，提取扫视次数、扫视幅度、水平方向视觉搜索广度、头部水平转动角度标准差等驾驶人换道意图特征指标， 构建了Logistic模型，并经效度检验后应用于对驾驶人换道意图的识别。结果显示，基于Logistic模型的驾驶人换道意图识别方 法的识别成功率达到90.24%，与基于转向灯信号的驾驶人换道意图识别方法相比，具有明显的时序及成功率方面的优势。 关键词关键词 车道变换；意图识别；Logistic模型 中图分类号中图分类号 U471.15 文献标志码文献标志码 A doidoi 10.3981/j.issn.1000-7857.2014.14.011 LaneChangingIntentIdentificationBasedonLogistic RegressionModel 1 2 1 1 PENGJinshuan,FURui,SHAOYiming,XULei 1.ChongqingKeyLabofTrafficSystemSafetyinMountainCities,ChongqingJiaotongUniversity,Chongqing 400074,China 2.KeyLaboratoryofAutomotiveTransportationSafetyTechnology,MinistryofTransport;ChanganUniversity,Xian 710064, China AbstractAbstract Toreducetheriskoflanechanges,amethodforlanechangingintentidentificationisproposedbasedonthelogistic model.ByusingfaceLABvisualtrackingsystem,experimentswereconductedunderrealroadenvironmentforthepurposeofstudying driverslanechangingintentidentification.Onthebasisofthedriversfixationcharacteristicsoftherearviewmirrorsbeforelane changingoperation,thesizeofthetimewindowforlanechangingbehaviorisdetermined.Basedondifferenceanalysisofvisual characteristics between lane keeping and lane changing intent stages, saccade numbers, visual search width in the horizontal direction, saccade amplitude, and standard deviation of head rotation angles in the horizontal direction are selected as the characteristic indice to identify drivers lane changing intent. The logistic model is constructed based on the leaning samplescharacteristics.Themodelisappliedtothelanechangingintentidentificationprocessaftert