Robust Identification of Smart Foam Using Set Membership Estimation in a Model Error Modelling Framework

The aim of this paper is robust identification of smart foam, as an electroacoustictransducer, considering un modelled dynamics due to nonlinearities in behaviour atlow frequencies and measurement noise at high frequencies as existent uncertainties.Set membership estimation combined with model error modelling technique is usedwhere the approach is based on worst case scenario with unknown but boundeduncertainties. The outcome is a robust identified model which consists of a nominalmodel with its uncertainty bounds that fits exactly the H∞ robust control schemewhich has been utilized in active noise control in recent years. While the nominalmodel has the desired physical characteristics as cut-off frequency and the anticipatedslope and flatness before and after this frequency, respectively, it is maintained in theacceptably tight uncertainty upper and lower limits, thus validating the identificationprocedure. Looseness and tightness of uncertainty strip has also been discussedregarding nonlinearities and measurement noise in low and high frequency regions.Meanwhile the identified nominal model can also be utilized in non-robust noisecontrol methods due to its lower order, reflecting the advantage of the appliedidentification approach

Keywords:

Smart foam , , Set membership , Estimation , Model error , , Robust identification