Robust Data-Based Model Predictive Control for Nonlinear Constrained Systems⁎. Issue 20 (2018)

• Record Type: Journal Article
• Authors: Manzano, J.M.; Limon, D.; de la Peñ, D. Muñz; Calliess, J.
• Is Part Of: IFAC-PapersOnLine. Volume 51:Issue 20(2018)
• Page Start: 505
• Page End: 510
• Publication Date: 2018
• Subjects: MPC -- Data-based control -- Machine learning -- Input-to-state stability; Automatic control -- Periodicals; 629.805
• DOI: 10.1016/j.ifacol.2018.11.039
• Languages: English
• ISSNs: 2405-8963

Abstract

Abstract: This paper presents stabilizing Model Predictive Controllers (MPC) to be applied to black-box systems subject to constraints in the inputs and the outputs. The prediction model of the controllers is inferred from experimental data of the inputs and outputs of the plant. Using a nonparametric machine learning technique called SPKI, the estimated (possibly nonlinear) model function is provided. Based on this, a predictive controller with stability guaranteed by design is proposed. Robust stability and recursive feasibility is ensured by using tightened constraints in the optimisation problem but without adding a terminal constraint on the optimisation problem. The proposed predictive controller has been validated in a simulation case study.