Physical Resource Block clustering method for an OFDMA cognitive femtocell system

This paper deals with a heterogeneous Orthogonal Frequency Division Multiple Access (OFDMA) network where a user-deployed low power femtocell operates in the coverage area of a traditional macro-cell using the same frequency band. Femtocells represent a promising solution to increase the network capacity in next-generation wireless networks. However, the arising interference must be mitigated by means of suitable resource allocation policies. In particular, an OFDMA system allows a flexible usage of the resources that are organized in Physical Resource Blocks (PRBs). This flexibility can be fully exploited in a heterogeneous network if the femtocell base station knows how the PRBs are organized and grouped. This paper considers a cognitive femtocell base station that is able to sense the environment and identify the set of PRBs allocated to a given user by the macro-cell base station. A PRBs clustering method is proposed here. Initially, suitable inputs are derived and then provided to the K-means algorithm for a clustering refinement. The method proposed here is able to correctly gather together the PRBs of each user. Performance comparisons with a hierarchical clustering method is presented. The benefits of PRBs clustering on direction of arrival estimation are shown in order to prove the effectiveness of the proposed methods.