AR Model Identification Using Higher Order Statistics

This paper presents a new approach to identify the parameters of autoregressive (AR) model from the third order statistics of the output sequence. The observed signal may be corrupted by additive colored Gaussian or non-Gaussian noise. The system is driven by a zero-mean independent and identically distributed (i.i.d) non- Gaussian sequence. The simulation results confirm the good numerical conditioning of the algorithm and the improvement in performance with respect to existing methods.