The electromagnetic spectrum is becoming scarce as the usage of wireless communications devices becomes ubiquitous. Current research has shown that by using multiple antennas at each wireless device, significantly higher densities of users can share the electromagnetic spectrum without reducing the performance of individual users. Most existing research in this area however is based on idealized system models that are often unrealistic. This project studies the extent to which the promising performance gains of multi-antenna systems predicted using idealized network models apply to networks models with more realistic assumptions, and is an important bridge between existing theoretical understanding and future practical implementation of a technology that could potentially alleviate the spectral crowding problem significantly.
Most existing research on spatially-distributed wireless networks with multi-antenna nodes is based on idealized models such as uniformly random spatial distribution of users, uncorrelated channels between antennas, and the availability of accurate channel-state-information (CSI); assumptions which often do not hold in practice. The investigator is studying the impact of non-uniform node distributions, channel correlations, and inaccurate CSI on the performance of wireless networks with multi-antenna nodes. Specific questions addressed include 1) What spectral efficiencies can be achieved in such networks? 2) To what extent can increasing the number of antennas per user with user density maintain constant perlink data rates in such networks? 3) What are appropriate models for spatial node distribution, channel correlation, and CSI uncertainty for spatially distributed networks? Answers to these questions help improve existing understanding of the performance benefits of multiple-antenna systems in realistic systems and in particular, their potential to alleviate spectral crowding in practice.
