摘要
为探究不同类型车辆运动特征差异对路段整体交通流状态判别的影响,解决单一数据源导致判别结果精度不高的问题,针对多种类型车辆的浮动车数据存在的显著差异,提出了基于改进D-S(Dempster-Shafer)证据理论数据融合的路段单元交通状态判别方法。通过分析不同类型车辆的浮动车数据,探究其在速度分布、交通状态划分标准及样本量等方面存在的差异;针对D-S证据理论在融合高冲突信息时的失效问题,从修正数据源基本信任分配函数与优化合成规则两方面改进D-S证据,并据此构建路段单元交通状态判别模型。经实例验证发现:基于出租车、公交车、私家车单一浮动车数据的路段单元交通状态判别准确率分别为83.58%,70.15%,61.19%,利用传统D-S证据理论融合数据的交通状态判别准确率为85.07%,改进方法判别准确率为94.03%。这表明改进方法可有效融合高度冲突的多种浮动车数据,其交通状态判别准确率高于基于单一浮动车数据或传统D-S证据理论的判别方法。
In order to explore the influence of motion characteristic difference of different types of vehicles on discriminating overall traffic flow state of road sections,and solve the problem of low ac-curacy of discrimination results caused by single data source,a method of judging the traffic state of road unit based on improved D-S(Dempster-Shafer)evidence theory data fusion was proposed,in view of the significant differences in floating vehicle data of various types of vehicles.By analyzing the floating vehicle data of different types of vehicles,their differences in speed distribution,traffic state division standard and sample size were explored.Aiming at the failure of D-S evidence theory when fusing high conflict information,the D-S evidence theory was improved by modifying data source ba-sic trust distribution function and optimizing synthesis rule.A traffic state discrimination model of sec-tion unit was constructed based on this.Then,the model was verified by an example.The results showed that the accuracy rates of the traffic state discrimination based on the single floating vehicle da-ta of taxis,buses,and private cars were 83.58%,70.15%,and 61.19%respectively.The accuracy rate of traffic state discrimination using traditional D-S evidence theory fusion was 85.07%,and the dis-crimination accuracy rate of the improved method was 94.03%.It shows that the improved method can effectively integrate a variety of highly conflicting floating vehicle data,and its accuracy of traffic state discrimination is higher than that of single floating vehicle data or traditional D-S evidence meth-od.
作者
王玉婷
李静
蔡晓禹
WANG Yu-ting;LI Jing;CAI Xiao-yu(Chongqing Key Lab of Traffic System&Safety in Mountain Cities,Chongqing 400074,China;College of Traffic and Transportation,Chongqing Jiaotong University,Chongqing 400074,China)
出处
《交通运输研究》
2021年第6期31-39,共9页
Transport Research
基金
国家自然科学基金青年科学基金项目(61703064)
重庆市技术创新与应用示范专项社会民生类重点研发项目(cstc2018jscx-mszdX0085)。
关键词
数据融合
D-S证据理论
路段单元
交通状态判别
浮动车数据
车辆运动特征
data fusion
D-S(Dempster-Shafer)evidence theory
road section unit
traffic state discrimination
floating vehicle data
vehicle motion characteristics
