Investigation into the Effects of Channel Properties on Training-based MIMO Channel Estimation

Abstract

This paper reports on investigations into the effect of channel properties on training-based MIMO channel estimation. Here, the channel's properties are represented by eigenvalues of the complex channel correlation matrix. The influence of these eigen values is assessed for two training based channel estimation methods, Scaled Least Square (SLS) method and Minimum Mean Square Error (MMSE) method. It is shown that for a given transmitted power to noise ratio in the training mode, the performance of the two estimation methods is governed by the sum of eigen values of the channel correlation matrix. Simulation results support this conclusion.