摘要
为有效刻画未来智能网联环境下交通流微观跟驰行为,以更加精确地进行车辆的运动决策,建立了基于安全势场理论下的车辆跟驰模型。模型以势场理论为基础,首先阐述了交通环境中安全势场的客观性、普遍性以及可测性,然后通过引入加速度参数对既有安全势场模型进行改进,改进后的安全势场模型能够有效刻画出在不同速度、加速度值下车辆安全势场的变化趋势。在分析安全势场变化基础上,构建的车辆跟驰模型强化了加速度参数对车辆跟驰行为的影响,由于不同速度、加速度信息在智能网联环境下车辆可以实时获取,因此该模型可应用于未来智能网联环境中。此外,在模型参数标定过程中,通过对NGSIM数据进行筛选,得到含有较多减速停车以及启动加速状态的轨迹数据,共筛选得到412组NGSIM真实跟驰车对数据,并最终利用人工蜂群算法对该模型进行参数标定。为评估模型仿真效果,选择OVM模型、IDM模型与本文模型进行比较,并选取均方根误差RMSE和平均绝对百分误差MAPE为参数标定结果评价与验证的指标,结果表明,建立的基于安全势场理论的车辆跟驰模型具有良好的精度,适用于描述考虑加速度参数条件下的跟驰行为,可为今后智能网联环境下车辆微观驾驶安全决策、交通流中观安全势场分布、交通流宏观状态估计等奠定理论基础。
To effectively characterize the micro-car-following behavior of traffic flows in future connected and automated vehicle environment,in addition to making more accurate vehicle motion decisions,a new car-following model based on the safety potential field theory was established.In this study,the objectivity,universality,and measurability of the safety potential field in traffic environments were first expounded,and then a model that improves the existing potential field model by introducing acceleration parameters was proposed.This improved model can effectively describe the vehicle safety potential fields under different speed and acceleration values.Through an analysis of this safety potential field model,a new car-following model was developed to enhance the influence of the acceleration parameters on the car-following behavior.Because velocity and acceleration information can be acquired in real-time in connected and automated vehicle environments,this model can be applied to car-following behavior descriptions in future environments.In addition,in the calibration process of the model parameters,trajectory data with more stop-and-go states were obtained by screening the NGSIM(Next Generation Simulation)data.Finally,412 pairs of car-following data were selected from the NGSIM dataset and used to calibrate the parameters of the car-following model based on the artificial bee colony algorithm.To evaluate the simulation effect of this car-following model,the RMSE(Root Mean Squared Error)and MAPE(Mean Absolute Percentage Error)were selected as the indicators for validating the calibration results.In addition,the OVM(Optimal Velocity Model)and IDM(Intelligent Diver Model)models were selected to compare with the car-following model in this study.The results reveal that the car-following model based on the safety potential field theory exhibits a good accuracy and is suitable for describing the car-following behavior considering the acceleration parameters.The findings from this study are helpful in laying a theoretical foundation for vehicle micro-driving safety decision-making,mid-view safety potential field distribution,and macro-state estimation under the environment of connected and automated vehicles.
作者
李林恒
甘婧
曲栩
冒培培
冉斌
LI Lin-heng;GAN Jing;QU Xu;MAO Pei-pei;RAN Bin(School of Transportation,Southeast University,Nanjing 211189,Jiangsu,China;Institute on Internet of Mobility,Southeast University and University of Wisconsin-Madison,Southeast University,Nanjing 211189,Jiangsu,China;Jiangsu Key Laboratory of Urban ITS,Southeast University,Nanjing 211189,Jiangsu,China)
出处
《中国公路学报》
EI
CAS
CSCD
北大核心
2019年第12期76-87,共12页
China Journal of Highway and Transport
基金
国家重点研发计划项目(2018YFB1600600)
江苏省自然科学基金项目(BK20160685)
东南大学优秀博士学位论文培育基金项目(YBPY1928).
关键词
交通工程
安全势场
人工蜂群算法
跟驰模型
智能网联环境
traffic engineering
safety potential field
artificial bee colony algorithm
car-following model
connected and automated vehicle environment
作者简介
李林恒(1991-),男,江苏南京人,工学博士研究生,E-mail:leelinheng@seu.edu.cn;通讯作者:曲栩(1982-),男,山东青岛人,副教授,工学博士,E-mail:quxu@seu.edu.cn。
