Semi-Blind Interference Alignment Techniques for Small Cell Networks

This paper investigates the application of blind and semi-blind interference alignment methods in clustered small cell networks. First, we examine two simple straightforward extensions of blind interference alignment and observe that neither of them is uniformly superior under all user distribution. Then, we exploit the location information of the users and base stations (BSs) in the cluster to define the semi-blind interference alignment problem, whose objective function is to minimize the supersymbol length by grouping the users that can be served in the same time slot. We show that this problem is NP-hard and propose two heuristic centralized algorithms that can provide effective solutions. The first scheme, top-level semi-blind supersymbol design, employs a grouping indicator matrix to determine the supersymbol while the second scheme, ordered semi-blind supersymbol design, constructs the supersymbol in an ordered and greedy fashion. Finally, we propose a more decentralized scheme which does not have the location information of all users and BSs in the cluster, but the BS utilizes the long-term SINR values, obtained via feedback, of each user it serves. By numerical simulations, we show that the proposed semi-blind algorithms perform uniformly better than pure blind interference alignment schemes for any possible user distribution scenario.