| 992 | 6 | 278 |
| 下载次数 | 被引频次 | 阅读次数 |
车路协同和车联网的发展为车辆群体之间的协作控制提供了可能。本文关注的是在车联网环境下,自动驾驶车辆群体避让动态障碍物的问题,目标是实现在不损失车辆个体效益的同时,可以达到车辆群体系统最优。本文提出了一种基于深度强化学习算法(DQN)的自动驾驶车辆群体协作避让动态障碍物的模型。模型在学习过程中考虑了车辆的安全性、单个车辆和车辆群体的行驶效率,并加入了车辆的换道协作机制。仿真验证结果表明,与现有的非协作避障模型相比,该模型可以显著地提高整体交通效率,在非常拥堵、比较拥堵和自由流三种给定的不同交通流状态下,车辆行驶效率(车辆平均速度)分别提高5.26%、21.44%、10.38%,整体车流量分别提高8.22%、34.47%、0%。
Abstract：The rapid development of connected vehicle technology and vehicle infrastructure cooperative systems has provided the possibility of cooperative control of vehicle swarms to avoid obstacles. This study examines the problem of automated vehicle swarm avoidance of dynamic obstacles in connected vehicle environments. The goal is to achieve an optimal swarm system without losing individual vehicle benefits.This study proposes a cooperative dynamic obstacle avoidance model for the automated vehicle swarm based on deep reinforcement learning. The proposed model considers the efficiencies of both individual vehicle and the vehicle swarm in the learning process, and a cooperative lane-changing execution model is proposed to ensure optimal decision making. Simulations showed that this model can significantly improve the overall traffic efficiency as compared with existing non-cooperative obstacle avoidance models. Under three given traffic flow conditions, namely, very congested, comparatively congested, and free flow, the increases in vehicle efficiency(i. e., average vehicle speed) were 5.26%, 21.44%, and 10.38% respectively, and the increases in overall traffic flow were 8.22%, 34.47% and 0% respectively.
[1]胡晓伟,石腾跃,于璐,等.基于扩展技术接受度模型的共享自动驾驶汽车用户使用意愿研究[J].交通运输工程与信息学报, 2021, 19(3):1-12.
[2]齐航,夏嘉祺,王光超,等.考虑出行者习惯与利他性偏好的自动驾驶网约车使用意向模型[J].交通运输工程与信息学报, 2021, 19(2):1-10.
[3]徐永.基于满意度的多目标约束模糊控制规则库的建立及应用[J].交通运输工程与信息学报, 2013, 11(01):74-78.
[4] ELMI Z, EFE M?. Path planning using model predictive controller based on potential field for autonomous vehicles[C]//IEEE. Proceedings of the IECON 2018-44th Annual Conference of the IEEE Industrial Electronics Society. New York:IEEE, 2018.
[5] KATHIB O. Real-time obstacle avoidance for manipulators and mobile robots[C]//Proceedings 1985 IEEE International Conference on Robotics and Automation, St. Louis:IEEE,1985:500-505, doi:10.1109/ROBOT. 1985. 1087247.
[6]修彩靖,郭继瞬,梁伟强.自动驾驶避障策略研究;2020中国汽车工程学会年会暨展览会,中国上海,2020[C].
[7] LAVALLE S M. Rapidly-exploring random trees:a new tool for path planning[J]. Computer. Science Dept Oct.1998.
[8] MA L, XUE J, KAWABATA K, et al. Efficient samplingbased motion planning for on-road autonomous driving[J].IEEE Transactions on Intelligent Transportation Systems,2015, 16(4):1961-76.
[9]王道威,朱明富,刘慧.动态步长的RRT路径规划算法[J].计算机技术与发展, 2016, 26(3):105-107, 112.
[10]宋晓琳,周南,黄正瑜,等.改进RRT在汽车避障局部路径规划中的应用[J].湖南大学学报(自然科学版),2017, 44(4):30-37.
[11] HART P E, NILSSON N J, RAPHAEL B. A formal basis for the heuristic determination of minimum cost paths[J].IEEE transactions on Systems Science and Cybernetics,1972, 4(2):28-29.
[12]马静,王佳斌,张雪. A*算法在无人车路径规划中的应用[J].计算机技术与发展, 2016, 26(11):153-156.
[13] KOMETANI E, SASAKI T. A safety index for traffic with linear spacing[J]. Operations Research, 1959, 7(6):704-720.
[14] LIAN Y, YUN Z, HU L, et al. Longitudinal collision avoidance control of electric vehicles based on a new safety distance model and constrained-regenerativebraking-strength-continuity braking force distribution strategy[J]. IEEE Transactions on Vehicular Technology,2016, 65(6):4079-4094.
[15]曾德全,余卓平,张培志,等.三次B样条曲线的无人车避障轨迹规划[J].同济大学学报(自然科学版), 2019,47(S1):159-163.
[16] BAKKER L. Multi-agent deep reinforcement learning for automated Highway driving[D]. Delft:Delft university of technology, 2019.
[17]单麒源,张智豪,张耀心,等.基于SAC算法的矿山应急救援智能车快速避障控制[J].黑龙江科技大学学报,2021, 31(1):14-20.
[18]姬浩,徐寅峰,苏兵.基于城市清洁车作业行为的移动瓶颈建模与仿真[J].系统工程学报, 2016, 31(5):676-688.
[19] WU K, GULER S I. Estimating the impacts of transit signal priority on intersection operations:a moving bottleneck approach[J]. Transportation Research Part C:Emerging Technologies, 2019, 105(3):46-58.
[20]徐建闽,杨招波,马莹莹.面向移动瓶颈的高速公路流量控制模型研究[J].广西师范大学学报（自然科学版）,2020, 38(3):1-10.
[21] PIACENTINI G, GOATIN P, FERRARA A. Traffic control via moving bottleneck of coordinated vehicles[J].IFAC-Papers On Line, 2018, 51(9):13-18.
[22]?I?I?M, JOHANSSON K H. Traffic regulation via individually controlled automated vehicles:a cell transmission model approach[C]//IEEE. Proceedings of the 2018 21st International Conference on Intelligent Transportation Systems(ITSC), New York:IEEE, 2018.
[23] PIACENTINI G, FERRARA A, PAPAMICHAIL I, et al.Highway traffic control with moving bottlenecks of connected and automated vehicles for travel time reduction[C]//IEEE. Proceedings of the 2019 IEEE, 58th Conference on Decision and Control(CDC), New York:IEEE, 2019.
[24]?I?I?M, JOHANSSON K H. Stop-and-go wave dissipation using accumulated controlled moving bottlenecks in multi-class ctm framework[C]//IEEE.Proceedings of the 2019 IEEE 58th Conference on Decision and Control(CDC), New York:IEEE, 2019.
[25] LIARD T, STERN R, LAURA M, et al. Optimal driving strategies for traffic control with autonomous vehicles[C]//The 21rst IFAC World Congress, 2020.
[26] MNIH V, KAVUKCUOGLU K, SILVER D, et al.Human-level control through deep reinforcement learning[J]. Nature, 2015, 518(7540):529-533.
[27] SUTTON R S, BARTO A G. Introduction to reinforcement learning[M]. Cambridge:MIT Press, 1998.
[28] KOBER J, BAGNELL J A, PETERS J. Reinforcement learning in robotics:a survey[J]. The International Journal of Robotics Research, 2013, 32(11):1238-1274.
[29] KRAU?S. Towards a unified view of microscopic traffic flow theories[J]. IFAC Proceedings Volumes, 1997, 30(8):901-905.
[30] KRAU?S, WAGNER P, GAWRON C. Metastable states in a microscopic model of traffic flow[J]. Physical Review E, 1997, 55(5):5597.
[31] ERDMANN J. SUMO’s lane-changing model[C]//BEHRISCH M, WEBER M. Modeling Mobility with Open Data, Spring International Publishing Switzerland,2015:105-123.
[32] GIPPS P G. Behavioral car-following model for computer simulation[J]. Transport Research. 1981, 15(2):105-111.
[33] NAIK G, CHOUDHURY B, PARK J-M. IEEE 802. 11 bd&5G NR V2X:Evolution of radio access technologies for V2X communications[J]. IEEE Access, 2019, 7(70169-84).
[34] ZHOU H, XU W, CHEN J, et al. Evolutionary V2X technologies toward the internet of vehicles:challenges and opportunities[J]. IEEE Proceedings of the IEEE,2020, 108(2):308-323.
[35] MISHRA P K, KUMAR A, PANDEY S, et al. Hybrid resource allocation scheme in multi-hop device-to-device communication for 5G networks[J]. Wireless Personal Communications, 2018, 103(3):2553-2573.
[36]付智俊,郭启翔,何薇,等.基于前车意图识别的自动驾驶车辆实时避障换道策略研究[J].汽车电器, 2020,(12):1-7, 11.
[37]彭涛,刘兴亮,方锐,等.智能汽车高速换道避障安全车距仿真分析[J].汽车工程师, 2020(12):36-41.
[38] RICKERT M, NAGEL K, SCHRECKENBERG M, et al.Two lane traffic simulations using cellular automata[J].Physica A:Statistical Mechanics and its Applications,1996, 231(4):534-550.
基本信息:
DOI：10.19961/j.cnki.1672-4747.2021.04.025
中图分类号:U495;TP18;U463.6
引用信息:
[1]沈悦,陈璟,周子涵,等.车联网环境下自动驾驶车辆动态障碍物协作避让模型[J],2021,19(04):13-23.DOI:10.19961/j.cnki.1672-4747.2021.04.025.
基金信息:
国家自然科学基金项目（52172333）;; 中央高校基本科研业务费（2682021ZTPY010）
2021-04-20
2021
2021-05-19
2021
1