Matched-Decision AP Selection for User-Centric Cell-Free Massive MIMO Networks

This work proposes a scalable access point (AP) selection framework based on a competitive mechanism that operates in two stages. Initially, the user equipment (UE) connects to an intermediate AP cluster, where a matched-decision among the APs and UEs establishes the best connection in terms of large-scale fading and prevents UEs from being dropped. Then, UEs can expand their AP clusters by connecting to more APs. We compare our method with baseline solutions considering different precoding techniques in centralized and distributed network implementations. We propose three strategies to fine-tune the AP clusters, reducing the number of UEs per AP without compromising the spectral efficiency (SE) or improving energy efficiency. The simulations comprise a range of UEs and APs, the number of antennas per AP varies, and each AP can serve a limited number of UEs. The results show that our solution improves up to 163% the SE of the 95% likely UEs compared with baseline solutions and that the number of UEs each AP serves is crucial for improving SE. Additional results indicate that our solution allows APs to save processing to enhance SE and evidence that the number of APs affects the SE differently in each network implementation.