Access Point Clustering-based Pilot Assignment for Cell-free Massive MIMO

Abstract

This paper presents an access point (AP) clustering-based pilot assignment scheme for cell-free massive multiple-input multiple-output systems. We use a graph to represent the relative positions of APs in the environment and cluster the APs into groups of similar behavior. Next, we assign the users to groups based on the set of APs that serves the users. Since users in different groups experience low interference, the pilots can be reused in each group while keeping the pilot contamination small. The proposed method has the advantage that the user grouping has low complexity. Further, the pilot assignment is performed for each group individually, reducing the complexity as it does not require a full search. We evaluate the performance of the proposed scheme by running Monte-Carlo simulations.