Clustering Scheme for Asynchronous Scalable Cell-Free Massive MIMO with Partially Coherent Transmission

In this paper, we develop a scalable framework for asynchronous cell-free (CF) massive multiple-input multiple-output (mMIMO) systems with partially coherent transmission. To address the scalability issue, we exploit the dynamic cooperation cluster concept and propose a novel scalable clustering algorithm. Specifically, in the first stage, the users are connected to a subset of access points (APs), aiming to make the users and APs establish the best connection in terms of large-scale fading. Next, the connected APs with the same quantized phase shift are grouped into a set of coherent clusters. By AP clustering, the APs in the same coherent cluster transmit data coherently, while the APs in different clusters transmit data in a non-coherent fashion. This can effectively improve the spectral efficiency of the asynchronous CF mMIMO system while ensuring the requirement of scalability. Finally, numerical results verified the effectiveness of the proposed method.