考虑车辆交互信息的换道风格辨识  

Lane change style recognition considering vehicle interaction information

作　　者：马宽 邓超 MA Kuan;DENG Chao(School of Automobile and Traffic Engineering,Wuhan University of Science and Technology,Wuhan 430065,Hubei,China;Wuhan University of Science and Technology Intelligent Vehicle Engineering Research Institute,Wuhan 430065,Hubei,China)

机构地区：[1]武汉科技大学汽车与交通工程学院,湖北省武汉市430065 [2]武汉科技大学智能汽车工程研究院,湖北省武汉市430065

出　　处：《农业装备与车辆工程》2023年第9期7-10,共4页Agricultural Equipment & Vehicle Engineering

基　　金：国家自然科学基金青年科学基金项目(52002298);湖北省自然科学基金计划青年项目(2020CFB118);湖北省教育厅科学技术研究计划青年人才项目(Q20201107);教育部产学合作协同育人项目(202102580026)。

摘　　要：随着智能交通系统的不断升级,在行车辆之间的信息交互越来越频繁,换道风格的体现会被周围在行车辆所干扰,因此需要考虑更多指标,使换道风格的分类与辨识更加科学。选择NGSIM数据集中的i-80数据集作为研究对象,考虑到处于匝道上的车辆的换道必需性,筛选强制换道的车辆数据,提取车辆的运动参数以及换道的信息交互参数。发现采用k-means算法将驾驶风格分为3个类别时,可以得到最稳定的聚类效果。基于Weka仿真平台,采用序列最小优化(SMO)算法和逻辑模型树(LMT)算法对换道风格进行辨识。结果表明,LMT算法对于换道风格具有更高的识别精度(90.57%),模型可为存在交互信息的交通环境安全评价提供理论依据。With the continuous upgrading of the intelligent transportation system,the information interaction between the moving vehicles is more and more frequent,and the manifestation of the lane changing style will be disturbed by the surrounding vehicles.Therefore,more indicators need to be considered to make the classification and identification of lane changing styles more scientific.The i-80 data set in the NGSIM data set was selected as the research object.Considering the necessity of changing lanes of vehicles on the ramp,the data of vehicles forced to change lanes was screened,and the motion parameters of vehicles and the information interaction parameters of lane changing were extracted.It was found that the most stable clustering effect could be obtained when the k-means algorithm was used to classify the driving styles into three categories.Based on the Weka simulation platform,the sequence minimal optimization(SMO)algorithm and the logic model tree(LMT)algorithm were used to identify the lane changing style.The results showed that the LMT algorithm had higher recognition accuracy(90.57%)for the lane changing style,and the model could provide a theoretical basis for the safety evaluation of traffic environment with interactive information.

关 键 词：驾驶风格 NGSIM 强制换道 SMO算法 LMT算法 

分 类 号：U461.91[机械工程—车辆工程]