An Actor-Critic-Identifier Control Design for Increasing Energy Efficiency of Automated Electric Vehicles

Researchers have developed a novel AI control system that significantly improves electric vehicle energy efficiency. The system uses a neural network identifier coupled with an actor-critic reinforcement learning framework to generate optimal driving commands. In simulations, this 'actor-critic-identifier' architecture increased total energy recovery by 12.84% compared to traditional controllers. The method learns the complex mapping between control inputs and power consumption online, enabling more efficient driving without relying on pre-defined models.