UC San Diego

The use of multiple antennas at one or both ends of a communication link can improve both the capacity and reliability of the system in a fading environment. However, the performance of the system depends heavily on the inherent structure of the channel itself. In this dissertation we focus our attention on the characterization of the wireless channel, modeling the channel, and the practical implications of the channel knowledge on the transmission strategy. The first part of this dissertation considers single-input multiple-output (SIMO) transmission systems. For these systems, receiver diversity is available to improve the reliability of the communication link. The improvement available is governed by spatial correlation. This portion outlines the construction of a dual channel measurement system and channel measurements. Modeling is used to further understand the environment and the received power patterns. The second part of this dissertation focuses on multiple-input multiple-output (MIMO) transmission systems. MIMO channel models are introduced, along with a framework for comparing the predictive performances of the models. Two analytical models, the Kronecker model and the structured eigenbasis model, are examined in detail. The differences in how the underlying physics in the channel are captured by the models are discussed, highlighting the performance impact. Geometrical models are used to investigate the model representations of the correlation matrix and the impact of system parameters on the model structure. The final portion of this work focuses on practical transmission system considerations. The ability of analytical models to accurately predict the performance of two systems using M-QAM modulation is investigated. The performances of the two analytical models, the Kronecker and Weichselberger models, are compared against the predicted performance using measured data. Reduced rank modeling, and other strategies to reduce feedback overhead are discussed. Additionally, the issue of how to best use the channel, showing when spatial multiplexing is preferred for correlated channels, is explored. Finally, the time sensitivity of feedback on a system is discussed. The effect of delay is investigated and the predictive performance of the analytical models is explored