首页/论文/2005.08245

arXiv 2005.082452020-05-17eess.SPcs.AIcs.LG

Dampen the Stop-and-Go Traffic with Connected and Automated Vehicles -- A Deep Reinforcement Learning Approach

Department of Philosophy, UMass Lowell（马萨诸塞大学洛厄尔分校哲学系）

Liming Jiang, Yuanchang Xie, Danjue Chen, Tienan Li, Nicholas G. Evans

英文摘要：

Stop-and-go traffic poses many challenges to tranportation system, but its formation and mechanism are still under exploration.however, it has been proved that by introducing Connected Automated Vehicles(CAVs) with carefully designed controllers one could dampen the stop-and-go waves in the vehicle fleet. Instead of using analytical model, this study adopts reinforcement learning to control the behavior of CAV and put a single CAV at the 2nd position of a vehicle fleet with the purpose to dampen the speed oscillation from the fleet leader and help following human drivers adopt more smooth driving behavior. The result show that our controller could decrease the spped oscillation of the CAV by 54% and 8%-28% for those following human-driven vehicles. Significant fuel consumption savings are also observed. Additionally, the result suggest that CAVs may act as a traffic stabilizer if they choose to behave slightly altruistically.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。

阅读原文

阅读 PDF ↗

arXiv 页面 ↗HTML 原文 ↗

本文内容

英文摘要相关论文

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。

继续探索

相关论文

Stop-and-Go: Exploring Backdoor Attacks on Deep Reinforcement Learning-based Traffic Congestion Control Systems

New York University Tandon School of Engineering（纽约大学坦登工程学院）；New York University Abu Dhabi（纽约大学阿布扎比分校）

Recent work has shown that the introduction of autonomous vehicles (AVs) in traffic could help reduce traffic jams. Deep reinforcement learning methods demonstrate good performance in complex control problems, including autonomous vehicle control, and have been used in state-of-the-art AV controllers. However, deep neural networks (DNNs) render automated driving vulnerable to machine learning-based attacks. In this work, we explore the backdooring/trojanning of DRL-based AV controllers. We develop a trigger design methodology that is based on well-established principles of traffic physics. The…

2021-08-26cs.CRcs.LGcs.SY

面向执行器与通信延迟下混合交通队列安全巡航的物理增强残差策略学习（PERPL）

Physics Enhanced Residual Policy Learning (PERPL) for safety cruising in mixed traffic platooning under actuator and communication delay

提出PERPL框架，结合物理模型与RL残差策略，用于混合交通队列分散控制，在执行器与通信延迟下实现更小车头误差与更好振荡阻尼。

2024-09-23cs.AIeess.SP

应用强化学习优化交通信号灯周期

Applying Reinforcement Learning to Optimize Traffic Light Cycles

KAIST（韩国科学技术院）

本文提出用强化学习实时优化交通信号灯周期，并通过 Simulation Urban Mobility 模拟器训练 Deep Q-Network，实验显示紧急停车平均次数减少 44.16%，证明其能有效缓解拥堵并改善交通流。

2024-02-22cs.AIcs.LG

Flow: 一种用于混合自主性的模块化学习框架

Flow: A Modular Learning Framework for Mixed Autonomy Traffic

Laboratory for Information and Decision Systems, Massachusetts Institute of Technology（信息与决策实验室，麻省理工学院）；Institute of Data, Systems, and Society, Massachusetts Institute of Technology（数据、系统与社会研究所，麻省理工学院）；Department of Mechanical Engineering, University of California, Berkeley（机械工程系，加州大学伯克利分校）

本文提出了一种模块化学习框架，利用深度强化学习解决复杂交通动态问题，通过提高系统层面的速度，使学习到的控制法则在仅有4-7%的自动驾驶汽车参与度下，相比人类驾驶性能提升高达57%。此外，在单车道交通中，一个仅使用局部观测的小型神经网络控制法则能够消除拥堵现象，达到近最优性能。

2021-12-30cs.AIcs.ROcs.SY

通过强化学习分析无信号交叉口混合交通的行为

Analyzing Behaviors of Mixed Traffic via Reinforcement Learning at Unsignalized Intersections

本报告利用强化学习分析无信号交叉口机器人车辆与人类驾驶车辆的混合交通行为，重点考察多模态分布、相关性及影响排队长度的因果因素。

2023-11-21cs.CYcs.LGcs.RO

自调节汽车：自动化自由流道路网络的交通控制

Self-Regulating Cars: Automating Traffic Control in Free Flow Road Networks

Department of Computer Science（计算机科学系）；New York University（纽约大学）；Indian Institute of Technology Delhi（德里印度理工学院）

本文提出了一种基于强化学习的自调节汽车方法，通过动态调节车辆速度来优化通行能力和防止拥堵，无需新基础设施，结合经典交通流理论和微观模拟，在高保真度的PTV Vissim模拟器上实现了提高通行能力、减少延误和停车次数的改进。

2025-06-13cs.LG

更多相关论文收起更多论文

基于深度强化学习的智能交通信号控制与优化的CO2排放

Deep Reinforcement Learning-based Intelligent Traffic Signal Controls with Optimized CO2 emissions

Simon Fraser University（西蒙弗雷泽大学）；Breeze Traffic Inc.（微风交通公司）

针对交通信号控制中CO2排放研究不足的问题，提出强化学习奖励塑造方案EcoLight，在减少CO2排放的同时保持旅行时间等指标竞争力，并比较了Q-Learning、DQN、SARSA和A2C算法在多场景下的性能。

2023-10-23cs.AIcs.LGcs.SY

CAV-AHDV-CAV：通过新型跟驰结构与强化学习缓解网联自动驾驶车辆的交通振荡

CAV-AHDV-CAV: Mitigating Traffic Oscillations for CAVs through a Novel Car-Following Structure and Reinforcement Learning

针对混行交通中的跟驰交通振荡问题，提出CAV-AHDV-CAV跟驰框架与深度强化学习方法，经多数据集验证在避撞等方面优于基线，可提升交通效率与安全性。

2024-07-11cs.AIcs.RO

Vessel-following model for inland waterways based on deep reinforcement learning

Technische Universität Dresden（德累斯顿工业大学）

While deep reinforcement learning (RL) has been increasingly applied in designing car-following models in the last years, this study aims at investigating the feasibility of RL-based vehicle-following for complex vehicle dynamics and strong environmental disturbances. As a use case, we developed an inland waterways vessel-following model based on realistic vessel dynamics, which considers environmental influences, such as varying stream velocity and river profile. We extracted natural vessel behavior from anonymized AIS data to formulate a reward function that reflects a realistic driving styl…

2022-07-07cs.CEcs.LG

使用深度强化学习和真实轨迹数据的自动驾驶汽车交通平滑控制器

Traffic Smoothing Controllers for Autonomous Vehicles Using Deep Reinforcement Learning and Real-World Trajectory Data

UC Berkeley（加州大学伯克利分校）；École des Ponts ParisTech（巴黎高科路桥学校）

本文基于 I-24 真实轨迹数据，用深度强化学习训练仅依赖前车信息和可选下游信息的交通平滑巡航策略，在 4% 渗透率下实现超过 15% 的燃油节省，并验证换道与信息缺失下的鲁棒性。

2024-01-18cs.AIcs.MAcs.SY

跟车模型与环形道路下的拥堵控制：在已知延迟下考察极限环

Car-following Models and Congestion Control with Followerstopper on a Ring-Road under Known Delay -- Examining Limit Cycle

AI, Autonomy, Resilience, Control Lab (AARC Lab) Department of Electrical & Computer Engineering The University of Alabama in Huntsville, Huntsville, USA（人工智能、自主性、韧性、控制实验室（AARC实验室）电气与计算机工程系亚拉巴马大学亨茨维尔分校）

本文研究了在环形道路中使用Followerstopper控制车辆以消除拥堵波并提高交通稳定性，通过模拟展示了已知延迟对交通流的影响。

2025-12-03cs.ROcs.SYeess.SP

Connected Vehicle Supported Adaptive Traffic Control for Near-congested Condition in a Mixed Traffic Stream

Connected Vehicles (CVs) have the potential to significantly increase the safety, mobility, and environmental benefits of transportation applications. In this research, we have developed a real time adaptive traffic signal control algorithm that utilizes only CV data to compute the signal timing parameters for an urban arterial in the near congested condition. We have used a machine learning based short term traffic forecasting model to predict the overall traffic counts in CV based platoons. Using a multi objective optimization technique, we compute the green interval time for each intersecti…

2019-07-14eess.SP

Estimating Transit Vehicle Dwell Times at Bus Stops

This work presents a quantitative approach to estimate the total time spent in the vicinity of a bus stop including the deceleration time, the boarding and alighting time (developed in an earlier study), the acceleration time, and re-entry time (time required to merge into the adjacent lane). Different statistical models were used to compute the deceleration, acceleration and merge times. Typical deceleration and acceleration levels were computed using kinematic equations that were then used to compute both the deceleration and acceleration times. The adopted method to estimate both the decele…

2020-08-10eess.SPphysics.soc-ph