考虑智能网联车辆影响的八车道高速公路施工区可变限速控制方法  

Variable speed limit control method in work zone area ofeight-lane highway considering effects of connected automated vehicles

作　　者：过秀成[1] 肖哲 张一鸣 张叶平 许鹏宇 Guo Xiucheng;Xiao Zhe;Zhang Yiming;Zhang Yeping;Xu Pengyu(School of Transportation,Southeast University,Nanjing 211189,China;Institute of Transport Planning and Engineering,Jiangsu Provincial Planning and Design Group Co.,Ltd.,Nanjing 210019,China)

机构地区：[1]东南大学交通学院,南京211189 [2]江苏省规划设计集团有限公司交通规划与工程设计院,南京210019

出　　处：《东南大学学报（自然科学版）》2024年第2期353-359,共7页Journal of Southeast University：Natural Science Edition

基　　金：江苏省交通厅重大科技专项资助项目(2017Z02)。

摘　　要：为提升车联网环境下高速公路施工区交通运行效率及安全水平,提出了一种基于强化学习的可变限速控制方法.选取智能驾驶模型和真车试验模型,分别对传统人工车辆和智能网联车辆的跟驰行为进行建模,构建了以瓶颈下游路段交通流量为效率指标、瓶颈路段速度标准差为安全指标的复合奖励值,利用深度确定性策略梯度算法,分车道动态求解最佳限速值.仿真结果表明,所提可变限速控制方法在不同智能网联车辆渗漏率条件下均能有效提升交通流运行效率和安全水平,且在智能网联车辆渗漏率较低时,提升效果更加显著.当智能网联车辆渗漏率为1.0时,瓶颈下游路段交通流量提升10.1%,瓶颈路段速度标准差均值下降68.9%;当智能网联车辆渗漏率为0时,瓶颈下游路段交通流量提升20.7%,瓶颈路段速度标准差均值下降78.1%.智能网联车辆的引入能够提升至多52.0%的瓶颈下游路段交通流量.To improve the traffic operation efficiency and safety of highway work zone areas under the Internet of vehicles,a variable speed limit(VSL)control method based on reinforcement learning was proposed.The intelligent driving model and the model based on real vehicle experiments were selected to model the car-following behaviors of human-driven vehicles and connected automated vehicles(CAVs),respectively.A composite reward was constructed with the traffic throughput of bottleneck downstream segment(TTBDS)as an efficiency indicator and the standard deviation of bottleneck segment speed(SDBSS)as a safety indicator.The deep deterministic policy gradient algorithm was utilized to dynamically yield the optimal speed limit values for each lane.The simulation results show that the proposed VSL control method can effectively improve traffic flow efficiency and safety under different penetration rates(PRs)of CAVs.The improvement is more obvious under lower PRs of CAVs.When the PR of the CAVs is 1.0,the TTBDS is increased by 10.1%and the mean of SDBSS is decreased by 68.9%.When the PR of the CAVs is 0,the TTBDS is increased by 20.7%and the mean of SDBSS is decreased by 78.1%.The introduction of CAVs can increase the TTBDS by up to 52.0%.

关 键 词：可变限速控制 深度确定性策略梯度算法 八车道高速公路施工区 智能网联车辆 协同自适应巡航控制 

分 类 号：U491.54[交通运输工程—交通运输规划与管理]