Intellectual Merit: The generalized view of multiuser communication with side information resulting from the proposed activity will be a significant contribution to network information theory. It will set in place powerful tools and techniques that will be pivotal in estimating the capacity limits of complex wireless networks. Genie based capacity bounds will establish fundamental limits of how useful side information can be in a wireless network. The proposed activity brings together diverse areas like networking, optimization theory, and numerical algorithms. The PI has a successful record of previous research characterizing single and multiuser capacity with partial channel knowledge. Broader Impact: 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 results of this research will be disseminated broadly through traditional scholarly venues to the entire research community, the government and the industry. The research will help government agencies direct their resources into enabling the applications that are shown to be within the capacity limits of wireless networks. The guidelines for designing wireless networks will be useful for the industry. Society as a whole will benefit from the increased array of applications of wireless networks driven by the optimal design principles. The results of the proposed activity will be integrated into coursework and educational research initiatives. The fundamental nature of these contributions will allow simple abstractions that can be readily understood by undergraduates, inspiring them to seek careers in research and education. As a minority institution, UC Irvine allows many opportunities for the vision of future wireless networks emerging out of this research to be presented to minorities, women and disadvantaged sections of society.
Rapid growth in wireless networks combined with increased demands and potentially unlimited applications, make it necessary to understand the capacity limits of these networks. The problem is quite difficult because of the endless challenges and opportunities inherent in a wireless network. On the one hand we have limited bandwidth but on the other we have the potential to exploit spatial dimensions through cooperation and collaboration among nodes and by exploiting various forms of channel knowledge. Therefore, the opportunities for cooperation and availability of channel knowledge are the two major determinants of the capacity of wireless networks, and are studied in this project through the lens of communication with "side information". Our main findings are three-fold. First, we formulated the problem in classical information theoretic terms and arrive at a unified framework that not only combines previously known results into a common language but also allows us to build upon them, and to understand very basic questions such as the additional value added to the network capacity from a bit of "side information". The role of side information at the transmitters emerges as the key factor. Next we explored the capacity of cognitive radio networks. These networks have been proposed recently as a potential solution to the spectrum shortage problem. The enabling premise of cognitive radio networks is the side information available to the secondary users that allows them to co-exist with the primary spectrum users with minimal impact on the primary users' quality of service. Addressing the problem from an information theoretic perspective lead to new insights into distributed and dynamic communication, and elegant solutions to the question of how much spectrum sharing is optimal in cognitive radio networks. Finally, we explored the question of interference alignment with regard to the amount of channel knowledge. Interference alignment is a recent transformative and highly counter-intuitive idea that shows that wireless networks are not fundamentally interference limited. In their original form interference alignment schemes seem to require excessive amounts of channel knowledge, thus limiting their practical applications. However, in our studies we have found that perfect and instantaneous channel knowledge is not essential to interference alignment schemes and indeed one can achieve blind interference alignment. These findings fundamentally advance the state of art of network information theory, and have created much excitement in the information theory community as well as in the industry, leading to international research efforts aimed at further developing these ideas. The fundamental nature of these ideas has also enabled me to incorporate them into my teaching.
