Interference alignment-based sum capacity bounds for random dense Gaussian interference networks - Computer Science > Information Theory

Abstract: We consider a dense $K$ user Gaussian interference network formed by pairedtransmitters and receivers placed independently at random in a fixed spatialregion. Under natural conditions on the node position distributions and signalattenuation, we prove convergence in probability of the average per-usercapacity $\csum-K$ to $\half \ep \log1 + 2 \SNR$. The achievability resultfollows directly from results based on an interference alignment schemepresented in recent work of Nazer et al. Our main contribution comes through anupper bound, motivated by ideas of `bottleneck capacity- developed in recentwork of Jafar. By controlling the physical location of transmitter-receiverpairs, we can match a large proportion of these pairs to form so-called$\epsilon$-bottleneck links, with consequent control of the sum capacity.

Author: Oliver Johnson, Matthew Aldridge, Robert Piechocki

Source: https://arxiv.org/