Clustered Cell-Free Massive MIMO

This paper introduces a novel technique, clustered cell-free massive-MIMO ((CF)-F-2-M-MIMO), that generalizes the recently proposed cell-free massive MIMO (CF-M-MIMO) concept by optimizing the connectivity pattern among access points (APs) and mobile stations (MS's). Relying on the popular K-means clustering algorithm, APs and NISs are grouped together in clusters in such a way that strong interferers arising due to pilot contamination are minimized. The clustering pattern varies in accordance to the large-scale fading parameters and is therefore able to respond to macroscopic changes in the network (user mobility, network load variations). Numerical results show that the proposed architecture is able to support a large number of users even with very low-complexity processing at the AP side while greatly reducing the fronthaul capacity requirements.