A new Shifted Scaled LS channel estimator for Rician flat fading MIMO channel

In this paper, Training-Based Channel Estimation (TBCE) method is considered in the Rician flat fading Multiple-Input Multiple-Output (MIMO) channels. In this channel model, the performance of the conventional Least Squares (LS) channel estimator is probed first. A new shifted type of Scaled Least Squares (SLS) channel estimator, entitled as SSLS, is then proposed. It is a generalized form of the SLS technique. The optimal choice of training signals using Mean Square Error (MSE) criterion is also achieved. It is observed that the LS estimator cannot exploit the knowledge of channel statistics. However, the SSLS estimator exploits the trace of a specifically defined matrix of the channel covariance and the receiver noise power as well as the knowledge of first-order statistics about the channel. It is shown that in the Rician fading MIMO channel, the new channel estimator has better performance than the popular LS and SLS techniques. Theoretical analysis and simulation results demonstrate that increasing Rice factor is a reason for decreasing MSE of the proposed estimator.