Channel state prediction, feedback and scheduling for a multiuser MIMO-OFDM downlink

We consider the downlink of a MIMO-OFDM wireless systems where the base-station (BS) has M antennas and serves K ges M single-antenna user terminals (UTs). Users estimate their channel vectors from common downlink pilot symbols and feed back a prediction, which is used by the BS to compute the linear beamforming matrix for the next time slot and to select the users to be served according to the proportional fair scheduling (PFS) algorithm. We consider a realistic physical channel model used as a benchmark in standardization and some alternatives for the channel estimation and prediction scheme. We show that a parametric method based on ESPRIT is able to accurately predict the channel even for relatively high user mobility. However, there exists a class of channels characterized by large Doppler spread (high mobility) and clustered angular spread for which prediction is intrinsically difficult and all considered methods fail. We propose a modified PFS that take into account the ldquopredictabilityrdquo state of the UTs, and significantly outperform the classical PFS in the presence of prediction errors. The main conclusion of this work is that multiuser MIMO downlink yields very good performance even in the presence of high mobility users, provided that the non-predictable users are handled appropriately.