Description
Regulation of parameter-varying and nonlinear dynamics renders serious
problem in various fields of engineering from mechanical systems to
Information technology. Our small team has long been working on
elaborating a special branch of polytopic modelling techniques that
may leads to a powerful toolkit for solving real-world problems.
Polytopic modelling of parameter dependencies provides an opportunity
to design suboptimal state and output feedback. Furthermore, using
this technique the uncertainty of the models can also be taken into
account. The approach is based on Lyapunov’s indirect method and
during the last decades, various Lyapunov-function candidates were
elaborated according to the practical demands such as time-delay and
other specialties of the systems.
Through polytopic tensor-product models we are able to describe the
parameter dependencies separately, allowing to handle their properties
in a unified way independently of their measured/estimated or time
invariant nature. The goal is to construct methods that apply
different approaches for the parameter dependencies according to their
properties. Another important goal is the automatic determination of
these polytopic models from the LPV description.
The most important drawback of these methods compared to the
norm-bounded uncertainty techniques (and a great challenge for
mathematicians) is that - despite of the serious effort of the past
years – we have no efficient optimization method to solve robust
output feedback problem on polytopic models.
The talk is held in Hungarian!
Az előadás nyelve magyar!