Predictor performance of decision-directed channel estimation in 3GPP MIMO channels

OFDM-based multi antenna systems suffer from inaccurate channel estimates in time variant channels which lead eventually to performance degradation. Decision-directed channel estimation (DDCE) schemes have been proposed to reduce the number of reference symbols. A prediction of channel coefficients is necessary for MIMO detection, whose output is used to acquire new channel estimates recursively. Three approaches to predict channel coefficients are investigated in this work. For validation simulation were conducted with the 3GPP spatial channel model. Channel estimation and prediction accuracy for mobile terminals in a range of velocities and Signal to Noise Ratios were evaluated. Even though the predictors vary strong in computational complexity it is shown in this work the performance benefit of complicated approaches remains small in a variety of channel states.