Approximate ML detector for MIMO channels in unknown spatio-temporal colored noise with Kronecker product correlation

In this paper a new maximum likelihood (ML) based detector for multi-input multi-output (MIMO) channels in spatio-temporal colored noise fields is proposed. It is assumed a Kronecker model of spatio-temporal correlation of noise. Approximate ML (AML) detection algorithm of MIMO channels is considered for two cases: known noise correlation matrix and unknown noise correlation matrix. The ML decoder for the case of unknown correlation matrix is developed based on iterative procedure with successive estimation of symbols, spatial correlation matrix and temporal correlation matrix. The proposed method uses the Kronecker structure of spatio-temporal correlation matrix. Effectiveness of the proposed technique is confirmed by simulation results.