Cellular Interference Alignment (19 November 2015)

Interference alignment promises that in interference channels, each link can support half of a degree of freedom (DoF) per pair of transmit-receive antenna. However, this surprising theoretical result is achieved asymptotically with exponentially-long delay and/or exponentiallyhigh signal to noise ratio, with no gain in practical settings. Here, we aim to propose a communication scenario in wireless cellular systems, where the promised DoF gain of alignment is achieved without using time expansion or lattices. We focus on uplink cellular systems, where each cell is split into three sectors.

We assume that each sector receives interference from all four nearby out-of-cell sectors. For intra-cell sectors, we consider two cases: In the first case, we assume that sectors within one cell do not interfere (due to employment of directional antennas) and also do not collaborate in decoding. In the second scenario, we assume sectors within one cell do interfere, but receivers in each cell can collaborate in decoding. In addition, we assume that if one sector can decode a message, it can let all nearby sectors, out-of-cell or intra-cell, knows about the decoded message. 

In both cases, we propose a decoding order and alignment scheme which achieves one degree of freedom per two pair transmit-receive antennas with linear schemes without time expansion.We prove the result is optimum for all possible order of decoding and linear alignment schemes with no time expansion. To avoid the signaling schemes relying on strength of the interference link for communication which can cause unwanted noise enhancement, we introduce a notion of topologically robust scheme, which guarantees a minimum rate (or degrees of freedom) no matter if the interference link are strong or weak. In continue, we introduce a decoding order/alignment scheme for cellular systems which is topologically robust and still achieves the same optimum performance.