CIF: Small: Degree of Freedom Region of Interference Networks

Iowa State University, Ames IA

Investigators

Abstract

For point-to-point communications, channel capacity is very well understood, and there are schemes that approach the capacity for most useful channel models. In contrast, the multi-user communication channels are not as well understood. The channel models are more complicated, there are more varieties, and many important problems remain open. As one well-known example, the capacity problem for discrete memoryless broadcast channel is still unsettled. As another example, the interference channel capacity region is known only in certain special cases (such as strong interference). What is clear from the recent studies of such channels is that in a multi-user wireless network, the key performance limiting factor is often the interference rather than the noise, especially at high signal-to-noise ratio. To alleviate this difficulty, the notion of degrees of freedom has been used as a first order characterization of the system throughput as normalized by the logarithm of the signal to noise ratio. The total degrees of freedom of several interference networks have been quantified. On the other hand, the degrees of freedom region, which is more revealing than the total degrees of freedom, is much less well understood. For most of the interference networks, the regions remain unknown. This project seeks to quantify the degree of freedom regions of a class of general interference networks with multiple transmit nodes and multiple receive nodes, possibly equipped with multiple antennas, and with general message demands. In particular, we will study the X interference network, which include interference channels, broadcast channels, multiple access channels, as special cases. Through interference alignment design based on vector and rational independence, we will bound tightly the degrees of freedom region and devise schemes that are optimal (or near-optimal) in the degree-of-freedom sense. Broader Impact: The project will advance significantly our understanding of interference networks, and positively impact the design, architecture, and signal design and processing for next generation wireless networks. The project will also help train the next generation of researchers and engineers in the field of signal processing and communication networks.