Model-Based Reinforcement Learning for Eco-Driving Control of Electric Vehicles

Abstract

The optimal control problem of determining the speed profile of the vehicle for minimizing energy consumption is a challenging problem that necessitates the consideration of various aspects, such as the vehicle energy consumption, slope of the road, and driving environment. The proposed approach utilizes a model-based reinforcement learning algorithm that separates the vehicle's energy consumption approximation model and driving environment model. This allows for the utilization of domain knowledge of vehicle dynamics and the powertrain system for the reinforcement learning process, while maintaining model-free characteristics by updating the approximation model using experience replay. The paper also evaluates the proposed algorithm under various driving scenarios and compares it with other eco-driving control methods.

Tasks

Reproductions