Output prediction under random measurements. An LMI approach
| Authors: | Peñarrocha Ignacio, Universitat Jaume I, Spain
Sala Antonio, Universitat Politécnica de Valencia, Spain
Sanchis Roberto, Universitat Jaume I, Spain
Albertos Pedro, Universitat Politécnica de Valencia, Spain |
|---|
| Topic: | 1.1 Modelling, Identification & Signal Processing |
|---|
| Session: | Filtering and Estimation |
|---|
| Keywords: | missing-data, random sampling, unconventional sampling, time-varyingsampling periods, martingale convergence, linear matrixinequalities |
|---|
Abstract
In this paper, the design of an output predictor in a system withrandom scarce sampling is addressed. A model based predictor thattakes into account the past measured outputs is used, and aLyapunov function of the estimation error is used for designpurposes. The Lyapunov design problem becomes a feasibilityproblem over a set of linear matrix inequalities applying theSchur complement formula. Three different design approaches havebeen developed. Some examples show the performances of eachapproach.
