An eco-driving strategy for electric buses at signalized intersection with a bus stop based on energy consumption prediction

The interaction between signalized intersections and downstream bus stops often leads to increased energy consumption in electric city buses. To address this issue, this study proposes a specialized eco-driving strategy tailored for electric buses operating at signalized intersections and bus stops. First, this study investigates how driving behavior affects energy consumption and develops a predictive model for energy consumption utilizing real-world driving data. Second, an optimized reward function is formulated based on the constructed energy consumption prediction model, incorporating considerations of safety, efficiency, and comfort. Subsequently, an acceleration and deceleration strategy are established using the Soft Actor-Critic framework to generate an ecovelocity curve. The effectiveness of this strategy is evaluated against real-world driving data. When compared to the driving behaviors observed in three distinct real-world scenarios, the proposed strategy demonstrates energy savings of 31.19 %, 20.84 %, and 30.29 % for electric buses navigating signalized intersections and bus stops continuously.