Channel estimation for MIMO-OFDM systems in mobile wireless channels

A channel estimation method is proposed for multiple-input multiple-output orthogonal frequency di-vision multiplexing (MIMO-OFDM) systems in time-varying fading channels. In this method, a decision-direct-ed space-ahernating generalized expectation-maximization (SAGE) algorithm is introduced to the tracking of time-varying fading. In order to improve the estimation performance of the SAGE algorithm, a low rank approxi-mation method is presented by using the signal subspaee of the channel frequency autocorrelation matrix. The study reveals that this method can be incorporated into the SAGE algorithm. Furthermore, a modified fast sub-space tracking algorithm is given to adaptively estimate the signal subspace by utilizing training OFDM blocks sent at regular interval. Simulation results demonstrate the considerable benefits of the proposed channel estima-tion method.