Beamforming for MIMO-OFDM wireless systems

MIMO-OFDM is considered a very promising solution for present and future wireless communications because of its proven bandwidth efficiency and diversity and multiplexing benefits. Beamforming is one approach to achieve these benefits. In this work, we study beamforming in two contexts: quantized feedback beamforming and beamforming in the presence of interference.;We first study quantized feedback beamforming and compare three different quantized feedback schemes for a MIMO-OFDM system. Their performance is evaluated in terms of average error probability or packet error rate degradation and overhead. We consider different types of Doppler spectra, different values of delay spread, and different power allocation schemes, and we assess the benefits and restrictions of beamforming in these different scenarios.;A MIMO-OFDM system may suffer serious degradation in environments with high delay spread. In order to study this degradation, we first characterize the intercarrier interference and intersymbol interference using a per-tone processing model. Then a new ICI/ISI-aware beamforming scheme is proposed for use in a MIMO diversity system. We also propose suboptimal but efficient beamforming algorithms to achieve the beamforming gain while mitigating the interference.;In order to further improve throughput performance, an enhanced ICI/ISI-aware beamforming with a spatial multiplexing capability is proposed. Two schemes are studied. The first one iteratively optimizes the steering vectors to minimize the total mean square error of the estimation. The second one attempts to minimize the interference leakage and results in a closed-form solution with reduced complexity. In addition, a decision feedback equalizer is introduced to augment the interference suppression by reconstructing and subtracting the ISI based on the previous successfully decoded OFDM symbols. We show that all the ICI/ISI-aware schemes provide dramatic gains in performance when delay spreads are large.;By utilizing the degrees of freedom provided by multiple antennas, beamforming can also benefit multiuser MIMO systems, specifically, by reducing the interference among simultaneous users. For this purpose, we propose precoding/decoding techniques based on a new iterative zero-forcing (ZF) algorithm. This scheme optimizes the beamforming within the null space of the interference and achieves good sum-rate capacity and fast convergence. We show that it performs well compared to alternative schemes.;Finally, non-ZF beamforming approaches are studied to address the interference management in a MIMO backhaul network. Using an iterative generalized eigenvalue decomposition scheme, we attempt to optimize the sum-rate performance by minimizing the cross-talk among different user pairs. Compared with existing sum-rate optimization schemes, ours is more computationally efficient and suRTMers less multiuser interference. Using the proposed scheme, we generalize the study and provide guidelines and design strategies for beamforming in a MIMO backhaul network.