Transfer Learning-Based Multi-Objective Predictive Control for Vehicular Platoon

This paper proposes a novel transfer learning-based multi-objective predictive control (TLMPC) strategy to reduce the computational burden for the vehicular platoon. A data-driven model is established with subspace identification to alleviate the adverse effects of complex powertrain structure of electric vehicles. Then, a grey wolf optimizer is devised to navigate the search process, striving for globally optimal trade-offs among the conflicting multi-objectives. To improve the calculation efficiency, a transfer learning method is suggested to the platoon. Eventually, several experimental results demonstrate the effectiveness of the proposed TLMPC strategy.