Interference alignment: capacity bounds and practical algorithms for time-varying channels
Sidhom Farag, Michael George
MetadataShow full item record
Wireless communication systems are becoming essential to everyday life. Modern network deployments and protocols are struggling to keep up with these growing demands, due to interference between devices. The recent discovery of interference alignment has shown that, in principle, it may be possible to overcome this interference bottleneck in dense networks. However, most theoretical results are limited to very high signal-to-noise ratios (SNRs) and practical algorithms have only developed for interference alignment via multiple antennas. In this thesis, we develop new capacity bounds for the finite SNR regime by taking advantage of time-varying channel gains. We also explore practical algorithms for parallel single-antenna interference channels, which could arise due to orthogonal frequency-division multiplexing (OFDM). From the theoretical side, we study the phase-fading Gaussian interference channel. We approximate the capacity region in the very strong interference regime to within a constant gap. Our coding schemes combines ideas from ergodic and lattice interference alignment. On the practical side, we develop a matching algorithm for pairing together sub-channels for alignment. This algorithm relies on the concept of maximum weight matching from graph theory. Simulations demonstrate that this algorithm outperforms classical techniques when the network is interference limited.