GOPALA Kalyana

Thesis

Multiple antenna communications for 5G

 

“Multiple antenna communications for 5G”

Time Division Duplexing (TDD) Massive Multiple Input Multiple Output (MaMIMO) with a massive number of base station (BS) antennas relies on channel reciprocity to obtain Channel State Information at Transmitter (CSIT). However the overall end to end digital channel is not reciprocal due to the presence of Transmit (Tx) and Receive (Rx) chains which need to be corrected using calibration factors. Our work provides a simple and elegant expression of the Cramer Rao Bound (CRB) for calibration parameter estimation. We provide analysis for the existing least squares approaches and propose optimal algorithms to estimate the calibration parameters. We also consider beamforming for a rapidly time-varying point to point MIMO link. In an Orthogonal Frequency Division Multiplexing (OFDM) sytem, this results in inter-carrier interference (ICI). With an assumption of linear channel variation across the OFDM symbol, it is observed that the beamformer design problem is similar to that of a MIMO Interfering Broadcast Channel (IBC) beamforming design. The beamformer design takes into account receive windowing using the excess cyclic prefix and the window is jointly designed with the Tx beamformer. In addition to full CSIT, we also investigate partial CSIT approaches that maximize Expected Weighted Sum Rate (EWSR)  where the Tx has only partial knowledge of the channel. First, we use a large system approximation that also works well for a small number of Tx and Rx antennas to derive the beamformers. In our work, we also analyze the possibility of using the Expected-signal- expected-interference-WSR metric instead of the EWSR. Finally, experimental results on the Eurecom MaMIMO testbed are