Blind MIMO system identification based on cross-polyspectra

In this paper we propose a novel frequency domain approach for the identification of a multiple-input multiple-output (MIMO) system driven by temporally i.i.d. and spatially independent non-Gaussian processes. The system frequency response is obtained based on singular value decomposition of a matrix constructed based on the power-spectrum and slices of polyspectra of the system output. Since more than one HOS slices could be used, the SVD can be replaced by joint diagonalization of a set of matrices. The flexibility to select the HOS slices allows us to bypass the frequency dependent permutation and phase ambiguity problems, which are usually associated with frequency domain SVD.