Identification of errors-in-variables systems with ARMA observation noises

The purpose of the paper is to give the recursive estimate for coefficients of the single-input-single-output (SISO) errors-in-variables (EIV) system and to prove its strong consistency. The observation noises of the EIV system are assumed to be ARMA processes, and the system input is taken to be an iid sequence. The recursive estimation algorithm is given with the help of stochastic approximation (SA) method and the strong consistency is established under reasonable conditions. The numerical simulation is consistent with the theoretical analysis. (C) 2007 Elsevier B.V. All rights reserved.