Enhancing Energy Efficiency and Driveability in HEVs Through Real-Time Fuzzy Predictive Control

Electrified vehicles are more energy-efficient than their traditional counterparts. However, they require complex energy management strategies to achieve this improved efficiency. These strategies are not typically designed to solve the driveability and comfort problem caused by the drivetrain's backlash, which is exacerbated by the fast torque transient of electric motors. In this research, a real-time fuzzy adaptive predictive controller is proposed to enhance the energy efficiency, the driveability, and the comfort of parallel hybrid electric vehicles (HEVs). Firstly, a control-oriented prediction model for the parallel HEV is constructed, followed by the corresponding predictive energy management strategy. Secondly, a fuzzy logic module is integrated into the controller to identify the operating conditions of the electric motor and adapt in real-time a subset of the predictive controller's cost function weights. Finally, the tuning of the non-adaptive objective function weights is formulated as an offline optimization problem that is solved using the sequential quadratic programming (SQP) approach to achieve the best energy efficiency. Results from simulations and real-world experiments showcase significant improvements in energy efficiency, driveability, and comfort compared to state of the art methods, underscoring the merits of the proposed approach.