考虑异质交通流的随机参数优化速度跟驰模型  

Stochastic parameter-optimized car-following model considering heterogeneous traffic flow

作　　者：潘义勇[1] 全勇俊 管星宇 PAN Yiyong;QUAN Yongjun;and GUAN Xingyu(College of Automobile and Traffic Engineering,Nanjing Forestry University,Nanjing 210037,Jiangsu Province,P.R.China)

机构地区：[1]南京林业大学汽车与交通工程学院,江苏南京210037

出　　处：《深圳大学学报（理工版）》2024年第4期415-422,共8页Journal of Shenzhen University(Science and Engineering)

基　　金：国家自然科学基金资助项目(51508280);南京林业大学高学历人才基金资助项目(GXL2014031)。

摘　　要：为分析交通流异质性对车辆跟驰行为的影响,基于随机参数线性回归方法改进优化速度函数.根据分位数回归对交通流速度-密度数据进行分类,对每个类别数据进行随机参数线性回归,并得到不同类别的改进优化速度函数与假设检验结果,结合改进的优化速度函数和全速度差跟驰模型建立随机优化速度跟驰模型,利用傅里叶变化理论对跟驰模型进行稳定性分析,并搭建环形车道仿真平台对跟驰模型进行数值实验.结果表明,分类处理后的随机参数模型误差较未分类降低28%;随机参数跟驰车队的速度值随着0.5分位点车辆的增多而增大;随机参数跟驰模型车队较固定参数跟驰模型车队更能反映交通流异质性对车队的影响.建立的模型能够提高仿真维度,真实反映交通流的复杂运行状况.In order to examine the impact of traffic flow heterogeneity on vehicle following behavior,we propose an improved optimized speed function based on the stochastic parametric linear regression method.The speed-density data for traffic flow are categorized using quantile regression.Random parameter linear regression is then applied to each data category,resulting in improved optimal velocity function and hypothesis testing for each category.The stochastic optimal velocity car-following model is developed by integrating the enhanced optimal velocity function with full velocity difference car-following model.The stability of the car-following model is analyzed by applying Fourier transform theory.Numerical experiments on the car-following model are conducted through a simulation platform for circular lanes.The results indicate that categorization reduces the error of the random parameter model by 28%compared to the model without categorization.Additionally,the speed of the random parameter car-following fleet increases with the addition of 0.5 quantile vehicles.The random parameter car-following model fleet is better suited to reflect the impact of traffic flow heterogeneity on the fleet than the fixed parameter car-following model fleet.The model can enhance the simulation aspect and accurately depict the intricate functioning of traffic flow.

关 键 词：交通工程 交通流理论 分位数回归 随机参数线性回归 优化速度函数 跟驰模型 稳定性分析 

分 类 号：U491.1[交通运输工程—交通运输规划与管理]