Efficient Nonlinear Model Predictive Control by Leveraging Linear Parameter-Varying Embedding and Sequential Quadratic Programming

In this study, we are concerned with nonlinear model predictive control (NMPC) schemes that, through the linear parameter-varying (LPV) formulation, nonlinear systems can be embedded and with a sequential quadratic program (SQP) can provide efficient solutions for the NMPC. We revisit the different...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Karachalios, Dimitrios S, Abbas, Hossam S
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:In this study, we are concerned with nonlinear model predictive control (NMPC) schemes that, through the linear parameter-varying (LPV) formulation, nonlinear systems can be embedded and with a sequential quadratic program (SQP) can provide efficient solutions for the NMPC. We revisit the different constrained optimization formulations known as simultaneous and sequential approaches tailored with the LPV predictor, constituting, in general, a nonlinear program (NLP). The derived NLPs are represented through the Lagrangian formulation, which enforces the Karush-Kuhn-Tucker (KKT) optimality conditions for the optimization problem to be solvable. The main novelty suggests that the problem can still be efficiently solvable with a significantly lower computational load by approximating certain terms to reduce the computational burden. The proposed method is compared with other state-of-the-art approaches on standard performance measures. Moreover, we provide convergence analysis that can assert further theoretical guarantees in control, such as stability and recursive feasibility. Finally, the method is tested through well-studied control benchmarks such as the forced Van der Pol oscillator and the dynamic unicycle.
DOI:10.48550/arxiv.2403.19195