多车道复杂环境下汽车驾驶倾向性特征动态提取  

Dynamic Feature Extraction of the Vehicle-driving Propensity under Multi-lane Complex Conditions

作　　者：于翠翠[1] 王健[2] 王晓原[1] 王晓龙[1] 张敬磊[1] 孙亮[1] 

机构地区：[1]山东理工大学交通与车辆工程学院,山东淄博255091 [2]山东理工大学理学院,山东淄博255091

出　　处：《交通信息与安全》2014年第5期154-161,共8页Journal of Transport Information and Safety

基　　金：国家自然科学基金项目(批准号:61074140);山东省自然科学基金项目(批准号:ZR2010FM007;ZR2011EEM034)资助

摘　　要：对多车道复杂环境下汽车驾驶倾向性特征动态提取进行研究,有利于准确推演驾驶人意图,对汽车辅助(自动)驾驶系统、特别是其主动安全预警子系统构建也具有十分重要的意义。以3车道场景为例,分析目标车位于不同车道时周边的交通态势(主要指车辆集群编组关系,重点以目标车位于中间车道为例),设计实验采集人、车、环境等相关动态信息,运用粗糙集理论,进行基于最小信息熵的连续属性离散化和基于启发式贪心算法的属性约简,提取不同态势下驾驶倾向性类型特征向量。实验验证表明,文中提取的多车道复杂车辆编组关系下汽车驾驶倾向性动态特征向量是科学合理的,能够准确反映驾驶人倾向性。研究结果可为建立多车道复杂环境下驾驶倾向性动态辨识模型提供理论依据。To extract the dynamic feature of vehicle driving propensity under multi-lane complex condition is condu- cive to reason the drivers’ intent accurately, and has extreme significance for establishing the active driving （auto-driving） system, especially the active security warning system. With three-lane condition as an example, the traffic situation （con- sidering the vehicle groups, and focusing on the target vehicle in the middle lane here）, is analyzed when the target vehicle runs in different lanes. Experiments are designed to collect the dynamic information related to the driver, vehicle, envi- ronment and so on. Rough set theory is used to extract driving propensity type eigenvectors corresponding to different traffic situations based on continuous attributes diseretization of the minimum information entropy and attribute reduction of the heuristic greedy algorithm. The verification results show that, dynamic feature vector extraction of the vehicle-driv- er＇s propensity under multi-lane complex conditions is scientific and reasonable. The research provides a basis for the sub- sequent study of the dynamic recognition model of the drivels propensity adapted to the multi-lane environment.

关 键 词：交通安全 驾驶倾向性 粗糙集理论 车辆编组关系 动态特征提取 

分 类 号：U491[交通运输工程—交通运输规划与管理]