Eco-driving for Intelligent Electric Vehicles at Signalized Intersection: A Proximal Policy Optimization Approach

Konferenz: ISCTT 2021 - 6th International Conference on Information Science, Computer Technology and Transportation
26.11.2021 - 28.11.2021 in Xishuangbanna, China

Tagungsband: ISCTT 2021

Seiten: 7Sprache: EnglischTyp: PDF

Persönliche VDE-Mitglieder erhalten auf diesen Artikel 10% Rabatt

Zhang, Xiaoliang (Research Institute of Highway Ministry of Transport, Beijing, China)
Jiang, Xia; Xiong, Zhuang; Zhang, Jian (School of Transportation, Southeast University, Nanjing, China)
Li, Nan; Yang, Zhongyue (Jiangsu Expressway Network Operation & Management Co., Ltd., Nanjing, China)

The increasing number of motor vehicles promotes the consumption of transportation energy, which brings about a new research field named eco-driving. As a complex problem, the eco-driving operation near intersection is hard to be implemented. This paper utilizes the advanced reinforcement learning (RL) technique to solve the problem. A reward function that integrated safety, efficiency and energy conservation is designed, while a proximal policy optimization (PPO) algorithm is given to solve the continuous action space problem. Microscopic traffic simulations show that the smart electric vehicle can learn a speed profile that avoids unnecessary stopping near the intersection. Meanwhile, the proposed framework can significantly reduce the energy consumption of the electric vehicles by more than 66% with less increase of vehicle travel time. Moreover, sensitivity analysis shows that the trade-off between energy consumption and travel time can be flexibly adjusted by setting the parameters.