Resource Allocation Based on Clustered Interference Alignment with Redundant User Extraction in Ultra-Dense Networks

In this paper, we propose a resource allocation scheme with clustered interference alignment (IA) in Ultra-Dense Networks (UDNs). Our goal is to maximize the sum spectral efficiency (SE) of users according to the network distribution and provide additional subchannels for redundant users as well as ensure fairness of the primary users, while other works mainly consider the Degrees of Freedom (DoFs) demand of users, ignore how well one DoF can achieve and treat IA cluster as the object of subchannel allocation. Our method takes the coherence bandwidth as the minimum resource where precoding and equalizing of IA become effective in all resource blocks. We construct a weighted graph at first to cluster users with their corresponding small base stations into disjoint and different sizes of groups. Second phase, we utilize graph coloring to allocate primary subchannels to groups of users to achieve fairness. Then we propose a redundant user extraction scheme to allow users free of interference to join in primary subchannels of other cluster to increase sum SE of the network. Simulation results indicate in a certain transmit signal to noise ratio (transmit SNR) interval, the proposed solution has obvious advantages over existing scheme, promising for further research.