Multiuser channel estimation and prediction in two dimensions for MIMO–OFDM uplinks

This paper addresses the pilot-aided multiuser channel estimation for the uplink of multiple-input multiple-output (MIMO) orthogonal frequency division multiplexing (OFDM) systems. The channel estimate is performed in two dimensions, i.e., time and frequency (2D), by exploiting their correlation. With the properly designed separable pilot symbols, a multiuser signal separation can be achieved and a multiuser channel estimator can be designed as well. To reduce the high computational complexity for a truly 2D estimator, two cascaded one dimensional (1D) estimators are used instead, which corresponds to two estimation stages with two different filters. At the 1st stage, the Wiener filter is implemented in the frequency direction, which separates the received multiuser signals and outputs the tentative estimate corresponding to an individual user. Whereas, at the 2nd stage, the Kalman filter is adopted in the time direction to estimate and predict the desired channel state information.