Secure Topological Interference Management for Tactical Wireless Networks
Abstract
The pursuit of information dominance demands that tactical wireless networks be designed to operate as close to optimal as possible in all regards, from their data rates and security guarantees to their robustness to channel uncertainty. It is therefore especially important to understand the fundamental limits (information theoretic capacity) of these networks. While exact capacity characterizations remain elusive, progress has been made in the understanding of capacity approximations through Degrees of Freedom studies. These studies however, suffer from the severe limitation that they treat all non-zero channels as essentially equally strong, thereby neglecting to exploit the diversity of signal strengths that is present in any wireless network, i.e., the topological aspect of a wireless network. Topological interference management, as used in this proposal, refers to ways to exploit the finite precision knowledge of channel strengths at the transmitters. This is accomplished with the help of the Generalized Degrees of Freedom metric that is sensitive to both arbitrary channel strength levels as well as channel uncertainty levels. It is against this background that the proposed research seeks to study secure topological interference management. STATEMENT OF OBJECTIVES The study of secure topological interference management is motivated by three concerns: 1) Security, 2) Robustness, and 3) Optimality. In particular, the proposed research seeks information theoretic security guarantees that do not break down even with computationally unbounded adversaries, robustness to channel uncertainty ensured by allowing only finite precision knowledge of channel strengths to the transmitters, and information theoretic proof of optimality in the Generalized Degrees of Freedom (GDoF) sense that leads to approximately optimal schemes at finite SNR subject to further localized optimization around the GDoF optimal scheme. METHODS TO BE EMPLOYED The research space is comprised of three dimensions --- 1) the choice of metric (GDoF region leading to finite SNR rate region), 2) the choice of topology (weak, strong, mixed, symmetric or asymmetric), and 3) the choice of network (interference, broadcast, multicast, MIMO/SISO, arbitrary numbers of eavesdroppers). Within this research space, achievable schemes will utilize recent advances in interference alignment, rate splitting, power control, optimal jammer design and ways to exploit common randomness across distributed transmitters. Proofs of optimality, or converse proofs, will rely on aligned images bounds, sum-set inequalities, and functional sub-modularity. SIGNIFICANCE TO ADVANCEMENT OF KNOWLEDGE There is much prior work on secure communications, robust jamming schemes, and GDoF optimality, but very little that studies all three in conjunction. Typically optimal schemes for secure communications are studied with the assumption of perfect channel knowledge, robust jamming schemes are studied without any claim of information theoretic optimality, and GDoF studies with channel uncertainty ignore security concerns. Therefore, the secure topological interference management problem is virtually unexplored. One of the main hurdles to this work in the past has been the lack of effective information theoretic bounding techniques under channel uncertainty. Upper bounds are important not only to prove impossibility results, but also to discover key insights into the ideas needed to achieve those bounds. However, the recent advances in the Aligned Images bounds have broken this barrier, for the first time putting this rich research avenue within reach. The proposed research is aimed at exploring this exciting opportunity.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jun 10, 2019
- Source ID
- W911NF1910344
Entities
People
- Syed Jafar
Organizations
- Army Contracting Command
- United States Army
- University of California, Irvine