Towards a Third Generation of Interference Alignment - Low Latency Interactive Directional Networking
Abstract
Objective: Interference Alignment (IA) is a transformative idea that is essential to approach thecapacity limits of wireless networks. A variety of interference alignment schemes have shown thatin theory the capacity limits of wireless networks are orders of magnitude higher than what ispossible through current state of art interference management schemes. Focused on asymptotictheoretical guarantees of optimality, the first generation of IA schemes relied on a number ofidealized assumptions, such as precise channel state information at the transmitters (CSIT),instantaneous feedback, unlimited channel diversity, and asymptotically high (signal to noisepower) SNR levels, which made them difficult to translate into practice in traditional wirelesssettings. The second generation of IAschemes focused on departing from idealized assumptionsbut still adhered to classical models of wireless networks with trivial connectivity assumptions,isotropic transmission, and unbounded latency. We envision future networks that are based ondirectional networking using intelligent reconfigurable antennas to provide interactivecommunication with low latency. This new paradigm opens the doorto the next wave of innovationin interference alignment the third generation of interference alignment that is the goal oftheproposed research.Technical Approach: The proposal is comprised of 4 key thrusts. The first thrust explores thepotential for new IAsolutions based on Intelligent Reconfigurable Antennas. Where we previouslyhad dumb reconfigurable antennas capable of switching between random isotropic modes wenow have intelligent reconfigurable antennas that can not only choose the direction of transmissionbut also take advantage of any available channel knowledge to choose optimal switching patterns.The second thrust explores the potential for shaping and exploiting network topology throughdirectional transmission. Where we previously had only convex connectivity patterns due toisotropic transmission models we now have the possibility of creating specialized non-convexconnectivity patterns due to directional antennas. These connectivity patterns can be optimized forvarious communication and/or computation tasks that arise in a tactical network. The third thrustis focused on low latency (short) multiuser codes for sporadic erasures. The goal of this thrust isto develop multiuser erasure codes for sporadic erasures that may result from transientmisalignments of beams due to fluctuating connectivity. The fourth research thrust exploresanother challenging direction interactive communication. Where we previously considered onewaycommunication, or even two-way communication with independent flows in each direction,we may now be able to exploit causal dependencies that arise naturally in interactivecommunication.Anticipated Outcomes and Significance to DoD: Success along these objectives will usher in a3rd generation of IA schemes that bring the theoretical advantages of IA close to practice, allowunprecedented spectrum efficiency along with low latency for interactive communication. As such,these advances are targeted toward the goalof information dominance that is critical to defenseoperations of the future.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- May 05, 2021
- Source ID
- N000142112386
Entities
People
- Syed Jafar
Organizations
- Office of Naval Research
- United States Navy
- University of California, Irvine