Number-Based Clustering With Power Level Difference for Composite Fading Channels in Cell-Free Networks

A cell-free (CF) network with cloud radio access and user-centric clustering is a good option for forming user serving access point (AP) clusters. However, the cluster size should be limited in terms of cooperation complexity and cooperation gain. This paper developed a stochastic geometry approach to evaluate the performance of number-based clustering (NBC) with power level difference (PLD) under composite fading channels in the CF network. The specific cluster size of a typical user can be determined, and the effects of different network parameters on association probability and average data rate are analyzed. The results show that considering shadowing in the association model and channel conditions may weaken the cooperative gain and probability.