摘要：AbstractThis paper investigates multi-objective optimization of electric vehicle (EV) based on features extracted from three driving styles, aiming at coordinating dynamic performance, ride comfort and energy efficiency. First, an unsupervised learning approach is used to clusters real-world driving data, obtaining three different driving styles. Then, the preferred performances under distinct driving styles are analyzed, and driving-style-oriented are determined. A model predictive controller is developed so as to handle the formulated multi-objective optimization problem. Simulations are carried out under with the developed controller and system models. Simulation results showed that the proposed controller could well coordinate the dynamic performance, ride comfort and energy efficiency of the 4IWDEV, validating the feasibility and effectiveness of the developed methodology and algorithms.
