System Identification using Subband Signal Processing

DAMIÁN MARELLI; MINYUE FU

Orlando

Conferencia; IEEE Conference on Decision and Control (CDC); 2001

Institute of Electrical and Electronics Engineers (IEEE)

The purpose of this paper is of twofold. First, we give atutorial on a new approach to system identification us-ing subband signal processing. Secondly, we study twotypes of subband identification schemes, one using criti-cal sampling and one using over-sampling. We comparethe behavior of the suhhand identification techniquenith the traditional ?full-hand? identification techniquein terms of the computational cost, asymptotic residualerror and convergence rate. This comparison is used todemonstrate the potential power of the subbaud tech-nique. We also consider design issues for subhand iden-tification, especially for the case with over-sampling.