On MM-Type Channel Estimation for MIMO OFDM Systems

In this paper, the general channel estimator for MIMO OFDM systems is developed, the MSE bound of the estimator is derived to examine the issue of pilot tone placement, and the MM principle together with theory of majorization is applied to channel estimation to reduce computational complexity. It's well known that EM-type algorithms are powerful tools to estimate channel parameters through iterative calculation, actually, every EM-type algorithm is a special case of the more general class of MM algorithms. The first M of MM stands for majorize (minorize) and the second M stands for minimization (maximization). To construct an EM-type algorithm, specifying the complete data, computing the conditional expectation and maximizing the conditional expectation analytically are skillful and computationally complex. In contrast, the MM algorithms developed from the inequalities are easier to be understood and calculated. In addition to constructing algorithms based on the MM principle we analyze the convergence property of the MM algorithms. Finally, the simulation results demonstrate the performance of MM-type channel estimation algorithms