Interference is believed to be the principal barrier to achieving higher data rates in wireless networks. The recent emergence of the idea of interference alignment has shown that the capacity of wireless networks can be significantly higher than prior estimates. The key idea is to design signals cleverly to occupy half the bandwidth in such a manner that they cast overlapping shadows at each receiver where they constitute interference, while they remain distinct at the receivers where they are desired. The surprising conclusion is that "everyone gets half the cake", i.e., regardless of the number of users competing for the same wireless spectrum, potentially every user is able to access half the bandwidth with no interference from other users. This research integrates the interference alignment perspective into the existing array of wireless network capacity results and uses the collective insights, tools and techniques to develop optimal interference management algorithms. The integrated view of interference management resulting from this research is a significant contribution to both the theory and practice of wireless network design. It contributes analytical insights that are pivotal in not only estimating the performance limits of wireless networks but also in finding efficient ways to approach these limits. Understanding the capabilities of wireless networks is essential for the industry, the academia, the government agencies and the society in general to have realistic expectations from the wireless networks of the future.
The investigators adopt a layered approach with the following three thrusts - (1) Capacity characterizations for elemental scenarios, (2) Generalized degrees of freedom characterizations for different classes of networks, and (3) Distributed interference management algorithms for large networks. At the microscopic level, the research explores the capacity of a parameterized continuum of elemental networks that capture the transitional regime between scenarios where interference alignment is capacity optimal and the classical strong interference, very strong interference and noisy interference cases. On a coarse but larger scale, capacity approximations are obtained in the form of generalized degrees of freedom characterizations for wireless networks, including the impact of relays, feedback and cooperation. The third stage takes a macroscopic view of wireless networks and combines the insights obtained from the capacity and degrees of freedom perspectives into distributed interference management algorithms. The stability of network management algorithms and their robustness to channel uncertainty is also investigated.
