A Dissipativity Formulation for Stability Analysis of Nonlinear and Parameter Dependent MPC

Publikationstyp
Conference Paper
Date Issued
2018-08-09
Sprache
English
Author(s)
Cisneros, Pablo S. G.  
Werner, Herbert  
Institut
Regelungstechnik E-14  
TORE-URI
http://hdl.handle.net/11420/2488
Journal
Proceedings of the American Control Conference  
Start Page
3894
End Page
3899
Citation
Proceedings of the American Control Conference (2018-June): 3894-3899 (2018-08-09)
Contribution to Conference
American Control Conference, ACC 2018  
Publisher DOI
10.23919/ACC.2018.8431417
Scopus ID
2-s2.0-85052569666
Model Predictive Control (MPC) with stability guarantees usually relies on terminal constraints; this not only increases computational complexity but might also compromise feasibility. Furthermore, a priori stability guarantees are often not available and the prediction horizon has to be set long enough to ensure the state is driven into the terminal region. In contrast, this paper presents analysis conditions to guarantee stability a priori for parameter dependent and nonlinear systems with no added constraints, using a Linear Parameter Varying (LPV) MPC control law. Stability is guaranteed using dissipativity arguments, given known conditions on the behavior of the nonlinearity arising from constrained MPC. When applied to nonlinear systems via quasi-LPV modeling, this result also gives a subset of the region of attraction, via the admissible set of parameters which can be mapped to a region in state space.