Adaptive blind equalization of MIMO wireless channels using coupled parallel estimators

In this paper, a novel approach in which parallel Kalman filters (KF) are coupled to parallel recursive least squares (RLS) estimators is proposed for adaptive blind equalization of multi-input multi-output (MIMO) channels. Using the inverse of a FIR polynomial matrix, a regression model is developed to formulate parallel RLS algorithms to estimate the unknown channel parameters. Conditions for existence of the inverse of a FIR polynomial matrix are investigated. A state-space representation is formulated to develop parallel Kalman filters (KF) to estimate the state from which the input signals can be recovered and separated. Then, the KF are coupled to the RLS algorithms to jointly recover the input signals and separate them, as well as to estimate the channel parameters blindly from the channel output measurements. The proposed approach is corroborated with a simulation example on adaptive blind equalization of MIMO channels. Simulation results show that the proposed approach is effective in recovering and separating the source signals, and estimating the channel parameters blindly from the channel output measurements.