Optimization of Energy-Saving Driving Strategy on Urban Ecological Road with Mixed Traffic Flows

Abstract:

This paper addresses energy-efficient driving strategies for autonomous connected vehicles on ecological roads under mixed traffic flow conditions. A wildlife passage scenario， which significantly impacts energy-saving driving， is extracted， and an application framework for wildlife passages within the Internet of Vehicles （IoV） is developed. A driving model for vehicles on ecological roads under the IoV environment is also constructed， utilizing dynamic programming for discretized analysis and state division. An energy-efficient driving model for vehicles within mixed traffic flow is optimized and established. The Q-learning algorithm is applied to optimize and solve the energy-saving driving model for a single vehicle. Based on the ecological roads in Shanghai， a simulation scenario considering the risk of wildlife crossing is created to validate the energy-saving driving strategies in the IoV environment. The results show that the proposed energy-saving strategy can reduce vehicle fuel consumption by 6 % to 11 %. Additionally， the energy-saving effect improves with increasing traffic density of vehicles， verifying both the reasonableness of the model and the effectiveness of the algorithm.

Key words:traffic engineering;ecological roads;Internet of Vehicles;energy-saving driving;reinforcement learning