To characterize the assignment of mixed traffic flow composed of autonomous and human-driven vehicles, a road capacity function incorporating four variables, saturated headway, CAV penetration rate, and CAV travel time cost conversion function was proposed. On this foundation, a mixed-mode traffic flow equilibrium assignment model was constructed, and a CAV-oriented static iterative weighting method was adopted for model solution. Optimal dispersion coefficients corresponding to CAV penetration rates of 20%, 40%, and 60% were calibrated using numerical test cases, and a piecewise function relating penetration rate and dispersion coefficient was derived. The dispersion coefficient is set to 0.060 for penetration rates within the interval [20%, 40%), 0.158 for [40%, 50%), and 0.264 for [50%, 60%].
高峰时段,交通拥堵通常达到最严重的程度,给人们出行带来了极大不便,也对城市的经济和环境造成了负面影响。而网联自动驾驶车辆(Connected and automated vehicles,CAV)可以彼此更近地行驶,对提高路段容量、提升交通系统效率和缓解交通拥堵状况带来了新的可能性。然而,由于自动驾驶技术的普及和应用还处于初级阶段,目前道路上的车辆仍然以人工驾驶为主,自动驾驶车辆与人工驾驶车辆(Human driven vehicle,HDV)混合的情况会长期共存。因此,如何在高峰时段合理进行CAV和HDV混合交通流的分配,最大程度地提高道路通行效率,成为当前交通管理领域亟需解决的问题。
目前,针对传统交通流理论的研究较为全面,涵盖系统最优(System optimum,SO)、用户最优均衡(User equilibrium,UE)模型以及随机用户均衡(Stochastic user equilibrium,SUE)模型;但自动驾驶普及前的混合车流过渡期,融合CAV对路网通行能力、出行成本及流量分配的相关研究仍有待深化。
在国内外相关研究中,大量学者针对自动驾驶与人工驾驶混合分配模型的构建、求解以及数据仿真进行了深入探讨。在网联自动驾驶和传统人工驾驶混合交通流分配模型构建方面,Chen等[1]利用平衡模型和扩散模型,建立了一个随时间变化的部署模型,同时考虑了部署汽车车道带来的净收益。Li等[2]提出了结合自适应信号控制与网联自动驾驶车辆协同控制的深度强化学习方法,用于信号交叉口网联自动驾驶车辆与人工驾驶车辆混合交通的管理,结果表明,该方法在较低网联自动驾驶车辆渗透率下可降低旅行时间和燃油消耗。Guo等[3]提出了具有平衡约束的最优控制(Optimal control problems with equilibrium constraints,OCPEC)问题,其中HDV的路径选择行为由瞬时动态用户平衡(Instantaneous dynamic user equilibrium,IDUE)原则建模,而CAV的路径选择由上层的动态系统最优(Dynamic system optimum,DSO)原则建模。韩磊等[4]结合深度强化学习提出了混合交通流环境下基于改进竞争双深度Q网络的可变限速(Variable speed limit,VSL)控制策略。张佳翠[5]构建了双层模型,上层模型为系统出行费用最小,决策变量是CAV专用道部署位置和收费值,下层为混合交通均衡模型,通过变分不等式模型进行求解。
因此,本文针对高峰时段下自动驾驶与人工驾驶混合分配的问题,提出一种有效、合理的混合交通流分配模型,并对其进行深入研究和分析。首先,综合考虑道路网络混合车流4种不同情况的饱和车头时距,以及不同CAV渗透率进行路网通行能力分析。接着,考虑不同驾驶模式的不同出行时间成本函数,设定不同分散系数的Logit路径选择模型,构建高峰时段交通流分配模型。然后,基于该模型,提出混有CAV的静态连续平均(Method of successive averages,MSA)算法进行模型求解,即在每一步迭代前进行CAV和HDV混合交通流分配,而不是先将CAV迭代平衡分配后再迭代分配HDV。最后,通过对相关参数自动驾驶渗透率、分散系数、CAV出行时间成本折减系数进行深度分析,确定参数之间的最优对应关系以及各参数对于路网出行总成本的影响,以期为提高交通系统的运行效率和优化交通资源的分配提供理论支持。
表2为目前常用的启发式非数值算法的优缺点,对于混合用户均衡问题,计算其下降方向和迭代步长通常是困难的。因为每次的迭代过程都要通过交通流量的随机分配来确定目标函数的下降方向;而且,迭代的步长不像用户均衡问题是通过求解一维的极值问题计算得到最优解。因此,在实际中常用迭代加权法(Method of successive averages,MSA)对随机用户均衡问题求解。迭代加权法是一种介于增量分配法和平衡分配法之间的循环分配算法。
ChenZ B, HeF, ZhangL H, et al. Optimal deployment of autonomous vehicle lanes with endogenous market penetration[J]. Transportation Research Part C, 2016, 72: 143-156.
[2]
LiD, ZhuF, WuJ, et al. Managing mixed traffic at signalized intersections: An adaptive signal control and CAV coordination system based on deep reinforcement learning[J]. Expert Systems with Applications, 2024, 238(Part C): No.121959.
[3]
GuoQ Q, BanX G, AzizA H. Mixed traffic flow of human driven vehicles and automated vehicles on dynamic transportation networks[J]. Transportation Research Part C, 2021, 128: No.103159.
HanLei, ZhangLun, GuoWei-an. Variable speed limit control strategy based on improved reinforcement learning in mixed traffic flow environment[J]. Transportation Systems Engineering and Information, 2023, 23(3): 110-122.
[6]
张佳翠. 混合行驶环境下城市出行管理[D]. 北京: 北京交通大学交通运输学院, 2022.
[7]
ZhangJia-cui. Urban travel management in mixed driving environment[D]. Beijing: School of Traffic and Transportation, Beijing Jiaotong University, 2022.
[8]
BarzegariV, EdrisiA, NourinejadM. Fleet cost and capacity effects of automated vehicles in mixed traffic networks: a system optimal assignment problem[J]. Transportation Research Part C: Emerging Technologies, 2023, 148: No.104020.
GuoXin-gang, WangSong, ChengChao, et al. Cascading failures of mixed traffic flows in road networks considering critical density[J]. Journal of Jilin University (Engineering and Technoogy Edition), 2025,55(12): 3875-3884.
[11]
WangJ, PeetaS, HeX Z. Multiclass traffic assignment model for mixed traffic flow of human-driven vehicles and connected and autonomous vehicles[J].Transportation Research Part B: Methodological,2019, 126: 139-168.
[12]
ZhangF, LuJ, HuX J. Traffic equilibrium for mixed traffic flows of human-driven vehicles and connected and autonomous vehicles in transportation networks under tradable credit scheme[J]. Journal of Advanced Transportation, 2020(1): 1-18.
[13]
WangJ, WangW, RenG, et al. Worst-case traffic assignment model for mixed traffic flow of human-driven vehicles and connected and autonomous vehicles by factoring in the uncertain link capacity[J]. Transportation Research Part C, 2022, 140: No. 103703.
HaoWei, ZhangZhao-lei, LvNeng-chao, et al. A mixed traffic assignment model considering autonomous driving lanes[J]. China Journal of Highway and Transport, 2023, 36(10): 317-327.
[16]
ZhouZ M, YuanJ B, ZhouS M, et al. Characteristics analysis and equilibrium optimization of mixed traffic flow considering connected automated and human-driven vehicles[J]. Journal of Advanced Transportation, 2022(11): 1-15.
[17]
WangG F, JiaH F, TaoF, et al. Modelling the dual dynamic traffic flow evolution with information perception differences between human-driven vehicles and connected autonomous vehicles[J]. Physical A: Statistical Mechanics and its Applications, 2024, 640:No.129667.
[18]
WangJ W, ZhenG Y, XuQ, et al. Controllability analysis and optimal control of mixed traffic flow with human-driven and autonomous vehicles[J]. IEEE Transactions on Intelligent Transportation Systems, 2020(99): 1-15.
[19]
CuiS H, CaoF, YuBin, et al. Modeling heterogeneous traffic mixing regular, connected, and connected-autonomous vehicles under connected environment[J]. IEEE Transactions on Intelligent Transportation Systems, 2022, 23(7): 8579-8594.
[20]
MaL J, QuS R, RenJ, et al. Mixed traffic flow of human-driven vehicles and connected autonomous vehicles: string stability and fundamental diagram[J]. Mathematical Biosciences and Engineering, 2023, 20(2): 2280-2295.
[21]
MilanesV, ShladoverS E. Modeling cooperative and autonomous adaptive cruise control dynamic responses using experimental data[J].Transportation Research Part C: Emerging Technologies, 2014, 48: 285-300.
[22]
WangM, DaamenW, HoogendoornS P,et al. Potential impacts of ecological adaptive cruise control systems on traffic and environment[J]. IET Intelligent Transport Systems, 2015, 9(4): 352-358.
FagnantD J, KockelmanK. Preparing a nation for autonomous vehicles: opportunities, barriers and policy recommendations[J]. Transportation Research Part A: Policy and Practice, 2015, 77: 167-181.