Multiple-input-multiple-output (MIMO) wireless systems use multiple antenna elements at the transmitter and receiver to offer improved spectral efficiency over traditional single antenna systems. In these systems, properties of the transmit and receive antenna arrays play a key role in determining the overall performance of the system. This thesis derives an upper bound on ergodic (average) channel capacity which formally links good antenna diversity performance with good ergodic capacity. As a result of this derivation, antenna arrays with good ergodic capacity performance are designed in this thesis by designing antenna arrays with near-optimal diversity gain. Several approaches are developed to design antenna array elements which achieve near-optimal diversity. These design methods only require an array geometry and the power azimuth spectrum of the propagation environment. Examples and analysis are included that illustrate advantages and disadvantages of each design technique. Three different array geometries are also investigated. Diversity performance results for each design technique and array geometry, averaged over an ensemble of typical power azimuth spectrums, are presented and compared. This analysis shows that the diversity gain achieved by the best design approach is, on average, less than 1.5 dB below the optimal diversity gain.

This thesis proposes a method for determining the optimal antenna element radiation characteristics which maximize diversity gain given a specific power angular spectrum of the propagation environment. The method numerically constructs the eigenfunctions of the covariance operator for the scenario subject to constraints on the power radiated by each antenna as well as the level of supergain allowed in the solution. The optimal antenna characteristics are produced in terms of radiating current distributions along with their resulting radiation patterns. The results reveal that the optimal antennas can provide significantly more diversity gain than that provided by a simple practical design. Computational examples illustrate the effectiveness of adding additional elements to the optimal array and the relationship between aperture size or the description of the impinging field and the array performance. A synthesis procedure is proposed which uses genetic algorithm optimization to optimally place a reduced number of dipoles. The results from this procedure demonstrate that using the framework in conjunction with optimization strategies can lead to practical designs which perform well relative to the upper performance bound. Finally a novel array architecture is proposed where subsets of antennas are combined together into super-elements which are then combined in the same manner as the optimal array. The simplifications that result from the genetically optimized small array or the super-element array provide a design options which are feasible in many communication applications.