To address the issues of slow convergence and limited applicability of reinforcement learning models in automatic driving tasks, a two-tiered reinforcement learning framework is proposed as a substitute for the decision and control layers. Within this framework, the decision layer categorizes driving behaviors into lane keeping, left lane change, and right lane change. Subsequently, after the decision layer selects the appropriate behavior, execution is achieved by modifying the input to the control layer. Then, in combination with reinforcement learning and online experts, a new method RL_COE is proposed to train the control layer. Finally, the proposed algorithm is verified in the highway simulation environment based on Carla and compared with the baseline reinforcement learning algorithm. The results show that this method significantly improves the convergence and stability of the algorithm, and can better perform the driving task.
ZhangYu-xiang. Domain knowledge and physical-enhanced reinforcement learning for intelligent vehicles decision-making and control[D]. Changchun: School of Transportation, Jilin University, 2022.
[3]
SchwartingW, AlonsoM J, RusD. Planning and decision-making for autonomous vehicles[J]. Annual Review of Control, Robotics, and Autonomous Systems, 2018, 1: 187-210.
[4]
PadenB, ČápM, YongS Z, et al. A survey of motion planning and control techniques for self-driving urban vehicles[J]. IEEE Transactions on Intelligent Vehicles, 2016, 1(1): 33-55.
[5]
王景珂. 基于学习的自动驾驶行为决策研究[D]. 杭州: 浙江大学控制科学与工程学院, 2021.
[6]
WangJing-ke. Research on autonomous driving behavior decison-making based on learning[D]. Hangzhou: College of Control Science and Engineering, Zhejiang University, 2021.
[7]
SuttonR S, BartoA G. Reinforcement Learning: An Introduction[M]. Cambridge: MIT Press, 2018.
[8]
KendallA, HawkeJ, JanzD, et al. Learning to drive in a day[C]∥International Conference on Robotics and Automation, Montreal, Canada, 2019: 8248-8254.
[9]
WurmanP R, BarrettS, KawamotoK, et al. Outracing champion gran turismo drivers with deep reinforcement learning[J]. Nature, 2022, 602: 223-228.
LiWen-li, QiuFan-ke, LiaoDa-ming, et al. Highway lane change decision control model based on deep reinforcement learning[J]. Journal of Automotive Safety and Energy, 2022, 13(4): 750-759.
[12]
KiranB R, SobhI, TalpaertV, et al. Deep reinforcement learning for autonomous driving: a survey[J]. IEEE Transactions on Intelligent Transportation Systems, 2021, 23(6): 4909-4926.
[13]
ZhuM, WangY, PuZ, et al. Safe, efficient, and comfortable velocity control based on reinforcement learning for autonomous driving[J]. Transportation Research Part C: Emerging Technologies, 2020, 117: 102662.
[14]
HuangZ, WuJ, LyuC. Efficient deep reinforcement learning with imitative expert priors for autonomous driving[J]. IEEE Transactions on Neural Networks and Learning Systems, 2023, 34(10): 7391-7403.
[15]
MnihV, KavukcuogluK, SilverD, et al. Human-level control through deep reinforcement learning[J]. Nature, 2015, 518: 529-533.
[16]
HaarnojaT, ZhouA, AbbeelP, et al. Soft actor-critic: off-policy maximum entropy deep reinforcement learning with a stochastic actor[C]∥International Conference on Machine Learning, Stockholm, Sweden, 2018: 1861-1870.
[17]
LiangX, WangT, YangL, et al. Cirl: controllable imitative reinforcement learning for vision-based self-driving[C]∥Proceedings of the European Conference on Computer Vision, Munichi, Germany, 2018: 584-599.
[18]
HesterT, VecerikM, PietquinO, et al. Deep q-learning from demonstrations[C]∥Proceedings of the AAAI Conference on Artificial Intelligence, New Orleans, USA, 2018: 3223-3230.
[19]
ElallidB B, BenamarN, HafidA S, et al. A comprehensive survey on the application of deep and reinforcement learning approaches in autonomous driving[J]. Journal of King Saud University-Computer and Information Sciences, 2022, 34(9): 7366-7390.
[20]
KulkarniT D, NarasimhanK, SaeediA, et al. Hierarchical deep reinforcement learning: integrating temporal abstraction and intrinsic motivation[J]. Advances in Neural Information Processing Systems, 2016, 29: 990-998.
[21]
DuanJ, EbenL S, GuanY, et al. Hierarchical reinforcement learning for self‐driving decision‐making without reliance on labelled driving data[J]. IET Intelligent Transport Systems, 2020, 14(5): 297-305.
[22]
DosovitskiyA, RosG, CodevillaF, et al. CARLA: an open urban driving simulator[C]∥Conference on Robot Learning, Mountain View, USA, 2017: 1-16.