Game Theory Controller for Hybrid Electric Vehicles

This paper describes the development and experimental implementation of an energy management controller for hybrid electric vehicles (HEVs) based on the application of game theory (GT). This controller is constructed as a feedback Stackelberg equilibrium in the noncooperative game between the driver and the powertrain with the cost penalizing fuel consumption, NO_x emissions, battery state of charge deviation, and vehicle operating conditions deviation. This control policy is drive-cycle and time independent. A description of the controller implementation with ancillary strategy elements is given. Experimental results from tests in a parallel HEV prototype vehicle are presented and compared with the existing baseline controller in terms of fuel consumption and NO_x emissions. The HEV powertrain configuration is advanced and includes a high-speed diesel engine, two electric motors, and automated converterless transmission. Over the New European Driving Cycle (NEDC), the GT controller, with minimal calibration effort, demonstrates better performance than the existing baseline controller that is calibrated from the deterministic dynamic programing solution over NEDC. We also demonstrate that the GT controller substantially outperforms the baseline controller over other real-world-focused driving cycles while providing good drivability.