Abstract: Modern wireless networks are complex: as the number of users grows, not only does interference limit the performance, but the decentralized behavior of distributed terminals results in highly dynamical systems. Recent advances in network information theory lead to better fundamental understanding of interference. However, in the model it is usually assumed that interference is always present, which may not be the case in wireless data/ad-hoc networks, especially from the perspective of the physical layer. This is due to the intermittence and burstiness caused by the decentralized medium-access-control mechanism and/or networking protocols. Exploiting such burstiness can potentially yield significant gain in system capacity.
In this work we explore the benefit of feedback for interference management in bursty wireless networks. We investigate two different models of bursty interference channels (IC), where one model is motivated from the decentralized medium-access-control mechanism, and the other from the networking protocols that result in bursty traffic. For both models, we completely characterize the capacity region of the symmetric two-user linear deterministic bursty IC with feedback. We use insights from the deterministic case to give an approximate characterization of the symmetric capacity for the Gaussian bursty IC with feedback in the weak interference regime. Interestingly, feedback provides a synergic benefit in managing bursty interference, in the sense that it boosts up system capacity unboundedly in regimes where such gain is not achievable either in non-bursty networks with feedback or in bursty networks without feedback.
This is joint work with Changho Suh, Suhas Diggavi, and Pramod Viswanath.
Bio: I-Hsiang Wang received the Ph.D. degree in electrical engineering and computer sciences from University of California at Berkeley, USA, in 2011. From 2007 to 2011, he was affiliated with Prof. David Tse's group in Wireless Foundations, Berkeley. He is now a postdoctoral research associate in Prof. Christina Fragouli's group in École Polytechnique Fédérale de Lausanne (EPFL), Switzerland.
His research interests include network information theory, wireless networks, coding theory, and network coding.
Dr. Wang received a 2-year Vodafone Graduate Fellowship in 2006. He was a finalist of the Best Student Paper Award of IEEE International Symposium on Information Theory, 2011.