文档摘要：为研究驾驶特征指标与驾驶避险行为决策的关联规则以提升驾驶安全,将驾驶避险决策行为划分为纵向"停车避险"和横向"转向避险",并运用驾驶模拟系统构建12种山区公路交叉口交通冲突虚拟场景,招募38名驾驶入进行试验,采集车辆特征和驾驶入扫视、眨眼、注视等眼动特征数据.基于随机森林构建驾驶避险决策行为模型,然后引入沙普利加和解释(SHapleyAdditiveexPlanation,SHAP)框架进一步分析车辆、眼动特征与驾驶避险行为之间的非线性关系.结果表明:模型对纵、横向避险行为预测的准确率分别为84.77％、94.70％;纵向速度标准差、扫视持续时间标准差、轨迹偏差标准差、侧向速度标准差与驾驶避险决策行为存在明显关联,如纵向速度标准差过大(约大于10km/h),纵向"停车避险"可能性明显增加.
Abstract：Thisstudyaimstoimprovedrivingsafetybasedontheassociationrulesbetweendrivingcharacteristicindexandriskavoidancebehavior.Thisstudydividesdrivingriskavoidancestrategicdecisionbehaviorintotwotypes:longitudinal"brakingavoidance"andhorizontal"turningavoidance",andusesadrivingsimulationsystemtoconstruct12virtualscenesoftrafficconflictsatmountainousroadintersections.Theauthorrecruits38driversfordrivingsimulationtestsandcollectsvehiclecharacteristicsandeye-trackingdataincludingscanning,blinking,andgaze.Abehavioralmodelfordrivingavoidancedecisionsbasedonrandomforestsisconstructed.TheSHapleyAdditiveexPlanation(SHAP)frameworkisintroducedtofurtheranalyzethenonlinearrelationshipbetweenvehiclecharacteristics,eyemovementcharacteristics,anddrivingavoidancebehavior.Theresultsofthestrategicdecisionshowthattheaccuracyofthepredictionforthelongitudinalavoidancebehaviormodelis84.77％andtheaccuracyofthepredictionforthehorizontalavoidancebehaviormodelis94.70％.TheanalysisresultsfromtheSHAPframeworkshowthatlongitudinalspeedstandarddeviation,sweepdurationstandarddeviation,trajectorydeviationstandarddeviation,andlateralspeedstandarddeviationaresignificantlyassociatedwithdrivingavoidancestrategicdecisionbehavior.Forexample,whenthestandarddeviationoflongitudinalspeedisapproximatelygreaterthan10km/h,SHAP＞0.Mountainroadintersectionscanbepotentiallyriskyanddriverstendtomakebrakingavoidancedecisions;Whenthestandarddeviationoftrajectorydeviationisapproximatelygreaterthan0.45m,SHAP＞0.Driversareverylikelytoencountertrafficrisksatmountainroadintersectionsandneedtochangethevehicletrajectorytoavoidacollision,atwhichtimedriverstendtotakeaturningriskavoidancedecision.Byanalyzingthecharacteristiceffectof4samples,itisconcludedthatthecharacteristicindexesshowninthemodelcanwellreflectthestrategicdecisionbehaviorofdrivingavoidance.Accordingtotheassociationrulesofthecharacteristicindicators,thisstudycananticipateupcomingrisksatmountainroadintersectionsandavoidthemasmuchaspossible.
作者：秦雅琴 包丽馨 陈亮 勾钰 王锦锐Author：QINYaqin BAOLixin CHENLiang GOUYu WANGJinrui
作者单位：昆明理工大学交通工程学院,昆明650500
刊名：安全与环境学报 ISTICPKU
Journal：JournalofSafetyandEnvironment
年，卷(期)：2024, 24(6)
分类号：X951
关键词：安全工程 驾驶避险 决策行为 驾驶模拟 分类预测 沙普利加和解释(SHAP)
Keywords：safetyengineering drivingriskavoidance strategicdecisionbehavior drivingsimulation classificationprediction SHapleyAdditiveexPlanation(SHAP)
机标分类号：K091U491TP391.9
在线出版日期：2024年7月4日
基金项目：国家自然科学基金山区公路交叉口驾驶避险决策行为特性分析[
期刊论文] 安全与环境学报--2024, 24(6)秦雅琴 包丽馨 陈亮 勾钰 王锦锐为研究驾驶特征指标与驾驶避险行为决策的关联规则以提升驾驶安全,将驾驶避险决策行为划分为纵向"停车避险"和横向"转向避险",并运用驾驶模拟系统构建12种山区公路交叉口交通冲突虚拟场景,招募38名驾驶入进行试验,采集车...参考文献和引证文献
参考文献
引证文献
本文读者也读过
相似文献
相关博文
关键词：安全工程,驾驶避险,决策行为,驾驶模拟,分类预测,沙普利加和解释(SHAP),
|
|
1
