BLIND IDENTIFICATION IN NOISY ENVIRONMENT OF NONMINIMUM PHASE FINITE IMPULSE RESPONSE (FIR) SYSTEM USING HIGHER ORDER STATISTICS

摘要

The problem of estimating the parameters of nonminimum phase (NMP) finite impulse response (FIR) system from the cumulants of output observation corrupted by a Gaussian, or colored Gaussian noise, is considered and an approach based on 3rd order cumulants is presented. The developed approach is tested and compared with a method proposed by Zhang for various models with NMP characteristic and with various value of signal to noise ratio (SNR). To illustrate the theoretical results for a real process, we use the proposed approach to search for a model able to predict (or simulate) the daily insolation measured in Marrakesh, Morocco.