×

Efficient non-linear model based predictive control. (English) Zbl 1010.93044

It is shown that the closed-loop paradigm of linear time-varying systems together with the augmentation of state and feasible invariant ellipsoidal sets can be carried over to the case of nonlinear model based predictive control (NMPC). The key to this development is the use of a quadratic form of the triangle inequality in conjunction with a quadratic Lipschitz condition. The former is used to extract as much information as possible from linear approximations, whereas the latter is used in bounding the error of approximation. The advantages of the approach are illustrated by means of two numerical examples, NMPC of a coupled-tank system and the NMPC of a fixed-rotor helicopter.

MSC:

93B51 Design techniques (robust design, computer-aided design, etc.)
93C10 Nonlinear systems in control theory
93B50 Synthesis problems
PDFBibTeX XMLCite
Full Text: DOI