MIMO-OFDM Channel Estimation with Eigenbeamforming and User-Specific Reference Signals

MIMO-OFDM systems with channel reciprocity allow spatial multiplexing using eigenbeamforming. However, eigenbeamforming requires singular value decomposition, which is phase ambiguous and causes the precoded channel to become phase discontinuous in frequency domain. This changes statistical properties of channel components, which degrades the performance of channel estimation. In this paper we propose phase unifying method to improve channel statistics and consequently decrease the error in channel estimation. In addition, we investigate the 2D-Wiener channel estimation with respect to different channel frequency correlation functions. The performance is evaluated using LTE link simulator.