Lattice Based Structuring to Combat Interference in Simple Wireless Networks

Prof. Shlomo Shamai (Shitz), Technion-Israel Institute of Technology

Tuesday February 28, 2012, 3:15 PM

Venue: ETF C 109

Abstract

In this talk we demonstrate the advantage of the inherent algebraic structure of lattice codes in simple wireless networks. First, we consider the uplink channel in a simple Wyner type cellular model, with adjacent out of cell interference. A compute-and-forward relaying is employed, where cell sites decode equations of the transmitted bits by exploiting the channel interference. The penalty for non-integer channel coefficients is mitigated by a superposition strategy, which in certain regimes outperforms decode-and-forward, compress-and-forward, and standard compute-and-forward.

The second part of the talk focuses on mitigation of interference in a distributed, relay assisted network. The model includes one transmitter communicating to a remote destination via strongly interfered relays, which forward messages to the destination through reliable links with finite capacities. We study different settings of distributed processing, and find the scaling (pre-log) behavior. The upper bounds are based on cut-set arguments, combined with a multiple letter bounding technique. The lower bounds demonstrate that lattice based structures are in some cases advantageous to random coding arguments. It is shown that in interesting cases, alleviating interference (in terms of scaling), requires forwarding a definite amount of information about the interference along with the desired message.