Car Following Model and Algorithm Design based on Reinforcement Learning
2021 ◽
Vol 2083
(3)
◽
pp. 032008
Keyword(s):
Abstract Based on reinforcement learning technology, this paper establishes a new driverless car following model. DQN algorithm and traffic simulator are mainly used to train the agent, and the following model is finally obtained. Under the precise and controllable experimental environment, the preset optimization targets can achieve the expected assumption and complete the following behavior. This study will contribute to the development of unmanned vehicles in the future.
2019 ◽
Vol 8
(3)
◽
pp. 8619-8622
2018 ◽
Vol 97
◽
pp. 348-368
◽
2021 ◽
Vol 1910
(1)
◽
pp. 012019