The full spatial covariance matrix of the multiple input multiple-output (MIMO) channel is an important quantity in channel modeling, communication system signal processing, and performance analysis, and therefore this matrix forms the heart of the research outlined in this dissertation. The work begins with an investigation of a generalized framework for computing the full MIMO spatial covariance based on the power angular spectrum (PAS) of the multipath field and the transmit and receive antenna element radiation patterns. For the case of uniform linear arrays and when the PAS clusters satisfy uniform, truncated Gaussian, or truncated Laplacian distributions, a series expansion is used to allow analytic evaluation of the required integrals in the formulation. The study also demonstrates the validity of some simplifying assumptions used to reduce the complexity of the covariance computation by applying the technique to ray tracing data as well as considers an analysis of the convergence properties of the series when computed using a finite number of terms. The insights and tools obtained from this covariance analysis are then used to develop a general approach for constructing MIMO transmit and receive beamforming vectors based on the full spatial covariance. While transmit and receive beamforming for the MIMO channel is a well-studied topic, when the transmit precoding is based on channel covariance information, developing near-optimal transmit and receive beamformers when the receiver is constrained to use linear processing remains an unsolved problem. This iterative beamforming algorithm presented here can accommodate different types of available channel information and receiver capabilities as well as either a sum power constraint or a per-antenna power constraint. While the latter is more realistic, construction of the optimal transmit precoder is less understood for this constraint. Simulation results based on measured channels demonstrate that the approach generates beamformer solutions whose performance rivals that achieved for an optimal nonlinear receiver architecture.



College and Department

Ira A. Fulton College of Engineering and Technology; Electrical and Computer Engineering



Date Submitted


Document Type





MIMO systems, correlation, array signal processing, beam steering, time-varying channels, cooperative systems, radio propagation