SAGE Algorithm for Semi-Blind Channel Estimation and Symbol Detection for STBC MIMO OFDM Systems

In this paper, a novel semi-blind channel estimation and symbol detection based on space-alternating generalized expectation-maximization (SAGE) algorithm are proposed for space time block coded (STBC) multiple input multiple output orthogonal frequency division multiplexing (MIMO-OFDM) systems. At the receiver, the transmitted signals of all users are added. By utilization of SAGE algorithm, this superimposed received signals are decomposed into their signal components. Then, the channel is estimated at pilot positions. After that, the channel estimation in the other positions are obtained by interpolation and symbol detection is done by utilizing the estimated channel. In SAGE algorithm, selection of initial value is very important for the convergence. We discuss the appropriate range for selection of initial value in this paper. Simulation results show that by increasing the number of users, the slope of bit error rate curve increases with same initial value and the channel estimation becomes worse by increasing the iteration of SAGE algorithm if the initial value is not selected in the proposed rang.