abstract
- Model Predictive Control (MPC) has been a popular control strategy over recent years due to the possibility of ensuring optimal control while dealing with constraints. However, Nonlinear Model Predictive Control (NMPC) suffers from high computational complexity, resulting in long optimization times. This has restricted NMPC to nonlinear systems that have slow dynamics. This approach presents an MPC strategy for nonlinear systems using qLPV representations. In this strategy, the nonlinear dynamics of the nonlinear system are represented by scheduling variables, which are state dependent. Afterward, an LPV state space of the nonlinear system is derived. The future values of the scheduling parameter are determined based on the optimal planned trajectory from the previous MPC iteration. To ensure stability, a stable state feedback controller is designed for a terminal region where asymptotical stability is guaranteed. The stable state feedback controller and the full feasibility of the control strategy are ensured by using Linear Matrix Inequalities based on Lyapunov stability conditions. Finally, the control strategy is tested in two different nonlinear systems with fast dynamics, the van der Pol¿s oscillator and a quarter vehicle nonlinear suspension system. The performance of the control strategy is compared against a NMPC control strategy and a state feedback controller. The results proved that the proposed qLPV-MPC control strategy performs similar to the NMPC optimal control strategy while requiring significantly less computation time, making it suitable for real-time implementations on fast nonlinear systems. © King Fahd University of Petroleum & Minerals 2025.