A tunable explicit nonlinear MPC for a vehicle considering improved path tracking performance and computational efficiency

Model predictive controller (MPC) is an advanced technique for vehicle control due to its prediction ability. Although MPC can cope with multiple optimization targets, a serious computational load arises due to the online optimization. In this paper, a novel tunable explicit nonlinear MPC (E-NMPC), which provides satisfactory path-following performance and computational efficiency is proposed for a vehicle in the NATO double lane change maneuver. The proposed controller is based on an artificial neural network (ANN), which is a useful tool to model complex systems. The control law is generated offline to reduce the computational burden. In addition, two novel tuning mechanisms, which are based on ANN and fuzzy logic are proposed to adjust E-NMPC weights considering improved trajectory tracking performance. The simulations showed that the path tracking performance and computational speed can be improved by using the proposed tunable E-NMPC, compared to the classical online-optimization based NMPC.