Iterative Training Signal Design for MIMO Multiuser Systems

Abstract

An iterative training sequence design scheme called Iterative SuperImposed training sequence design with Multiple Interferers, or ISIMI, is proposed for estimating MIMO channels with colored noise. The proposed approach decomposes the MIMO channel and design the training sequence on a per channel basis, thus making the use of sequential minimum mean-squared error (MMSE) estimator ideal for channel estimation. The time-multiplexed-superimposed training (TM-SIT) transmission format is also proposed to accommodate the different training sequences obtained via the proposed ISIMI method. The proposed approach does not use nonlinear optimization as utilized in previous literature, nor make any assumption about the lack of interdependence between the transmitter and receiver. The approach can be proven to converge to at least a local optimal solution and is shown consistently by Monte Carlo simulation to outperform previously proposed MMSE based approaches by 4 dB for 4x4 MIMO systems, respectively, in terms of MSE when the sequential MMSE estimator is used.