Clustering Method for Reducing Backhaul Signaling Requirements among CPUs in Cell-Free MIMO Multiple CPUs System

The DCC(Dynamic Cooperation Clustering) approach traditional adopted in existing cell-free MIMO systems is beneficial for establishing user-centric clustering in a single CPU cell-free(A single Central Processing Unit Cell-Free) MIMO system However, the presence of only a single CPU in a real environment leads to an increase in the signal processing computational complexity and infrastructure costs. This paper proposes a clustering algorithm for Cell-Free Multiple CPUs MIMO systems to reduce the backhaul signaling requirements among CPUs. The approach involves configuring the CPU connecting the AP (Access Point) for clustering in progress and the CPU serving candidate UE (User Equipment) as identical as possible, thereby reducing the backhaul signaling requirements among CPUs. Each AP adopts a centralized cell-free operation, where the CPU centrally designs precoders and relays encoding signals to UEs in the downlink. The precoder design utilizes a scalable precoder, P-MMSE (Partial-Minimum Mean Square Error), at the CPU. Furthermore, power allocation adopts a fractional power control algorithm commonly used in centralized cell-free operations. The proposed clustering algorithm in this paper can alleviate the backhaul signaling burden among CPUs compared to conventional methods while simultaneously ensuring similar Spectral Efficiency (SE) performance to the existing approach.