Scalable Cell-Free Massive MIMO with Fully Distributed Large-Scale Fading Decoding

Cell-free massive multiple-input multiple-output (MIMO) is a promising technology for future wireless communication systems, where a large number of access points (APs) serve the user equipments (UEs) in the network coherently. Different ways of operating the system are possible, ranging from a fully centralized approach, where all signal processing is performed at a central processing unit (CPU), to a fully distributed approach, where most of the signal processing is performed at the APs. This paper proposes new heuristics to select the large-scale fading decoding (LSFD) weights in a fully distributed cell-free massive MIMO system. Due to the analytical intractability of the system, we rely on extensive Monte-Carlo simulations to evaluate their performance compared to the other operation levels. Importantly, we consider a scalable setup, where the APs are grouped into user-centric clusters. Our results show that a fully distributed system, which locally computes the LSFD weights in each AP using our heuristics, is able to approach the performance of a partially distributed system, where the LSFD weights are selected optimally in a centralized manner, especially for deployments with multiple antennas per AP. Moreover, when considering scalability, the proposed improved fully distributed system achieves substantial gains in terms of spectral efficiency over small-cell systems. Hence, even though fully centralized cell-free systems still achieve the best performance, distributed systems are a viable alternative that can be more practical to implement.