Multiple-input-multiple-output blind system identification based on cross-polyspectra

In this paper we propose a novel frequency domain approach for the n/spl times/n MIMO identification problem. The proposed method is a subspace approach that estimates the parameters of a FIR MIMO system, driven by temporally i.i.d. and spatially independent non-Gaussian processes, using cross poly-spectra of any order and cross-power spectra of the system output. Extensions to the case of non-white input signals are also discussed. Simulation results are provided to illustrate the performance of the new algorithm.