DIR based clustering for interference alignment in Ultra-dense networks

Ultra-dense networks are regarded as a promising candidate for providing high data rate at low implementation cost. With a dense and irregular deployment of access points, spatial reuse will be improved, and varying cellular coverage also known as the amorphous cell is provided, which leads to a more complicated interference environment. In this paper, a novel interference coordination strategy, based on applying interference alignment (IA) in each disjoint sub-cluster of these amorphous cells, is proposed. A main merit is its ability to effectively identify and mitigate the dominant interference for each victim user. The coordinating base stations in the sub-clusters are selected based on the ratio between the dominant interference (DI) and the rest of the perceived interference, called the dominant interference ratio (DIR), and thus interference alignment adapts to the varying interference situation by controlling the overall high-level interference. How to efficiently form IA sub-clusters has been well studied by a modified k-means algorithm. Simulation results demonstrate significant performance gains (about 30 %) of the proposed clustering IA strategy, compared with the random clustering case, and information sharing overhead is greatly reduced.