Graph-Based Iterative Gaussian Detection with Soft Channel Estimation for MIMO Systems

Conventionally, the uncertainties of channel coefficients are neglected, that is the estimated values of channel coefficients are taken as the true values in the stage of data detection. In the communications community, it is still an open question how to take into account the channel uncertainty for data detection/decoding, especially in a low-complexity manner. In this paper, we propose a low-complexity receiver algorithm which utilizes soft channel information. Channel coefficients are treated as variables and estimated in an element-wise manner. Their uncertainties are represented by the variances. Instead of performing channel estimation and data detection in a separate manner, this algorithm does everything in one stage, i.e., channel estimation and data detection/decoding are carried out simultaneously over a general factor graph. The feasibility of this algorithm is verified by means of Monte-Carlo simulations both in bit error ratio (BER) and channel estimation mean squared error (MSE).