Feedback and Precoding Techniques for MIMO Wireless Communications

Abstract

Multiple-Input Multiple-Output (MIMO) communication has a big potential to increase the throughput and reliability of wireless communication systems. In order to exploit this potential, it is crucial to adapt transmission according to the Channel State Information (CSI). This thesis investigates selected problems on feedback, precoding and performance analysis in CSI based multi-antenna communication.Specifically, the thesis consists of three parts,  in which different research problems in single user, multi-user and multicell radio communication scenarios  are considered.

The first part of the thesis examines the utilization of the spatial and temporal correlations of a MIMO wireless channel for limited feedback of CSI using Givens rotations. Though feedback using Givens rotations is well known, existing schemes fail to utilize the correlations of the channel for efficient feedback design. The main reason for this is that the distributions of singular vectors of a correlated MIMO channel or their component Givens rotations have been unknown. In this thesis, models for the distributions of Givens rotations of a spatially correlated MIMO channel are proposed. The proposed models are then utilized for bit allocations and efficient codebook design for CSI feedback, which results in significant improvement in system performance. In addition, a new algorithm called Adaptive Rotative Quantization (ADRQ) is proposed for efficient feedback design based on adaptive utilization of the spatio-temporal correlations of the MIMO channel.

The second part of the thesis deals with the problem of equivalent channel (precoder) coordination in multi-user MIMO systems. In downlink multi-user MIMO, pre-coded reference signals or feed-forwarding is required to coordinate either equivalent channel information or precoder  information from the base station to the users. Pre-coded reference signals, especially user specific pilot symbols, have a large overhead and complexity; on the other hand, limited capacity feed-forward links cause performance degradation. In this thesis, a novel technique called Partial Equivalent Channel Inversion (PECI) is proposed. PECI based precoding avoids the need for coordination of equivalent channel information without a significant performance loss. The performance of the PECI technique is evaluated and a new adaptive mode selection algorithm is proposed to further improve the performance of PECI based precoding. It is also shown that the proposed technique has lower feedback requirement than conventional multi-user precoding schemes.

The third part of the thesis investigates the effect of CSI feedback and back-haul delays on the performance of cooperative multicell networks with multiple antenna base stations. Wireless systems are affected by CSI delay; however, the effect is worse in multicell systems in which CSI is exchanged through the back-haul network. In this thesis, performances of egoistic and altruistic beamforming strategiesin a multicell network are evaluated and compared in presence of  feedback delay, back-haul delay and limited feedback of CSI. Closed form expressions are obtained for approximations of the average sum rates and a switching criterion is subsequently proposed to improve performance by switching the transmission strategy based on the signal-to-noise ratio and feedback exchange properties. In order to improve the multicell performance further, a delay tolerant beamforming scheme is proposed based on weighted combinations of maximum ratio transmission and intercell interference cancellation beamforming schemes.