Abstract
In this work, we analyze the performance of the downlink of a cell-free massive multiple-input multiple-output (mMIMO) system considering finite capacity fronthaul links. We model the locations of the remote radio heads (RRHs) and the users as two independent binomial point processes (BPPs). Conditioned on the locations of the RRHs and users, and considering imperfect channel state information (CSI) and conjugate beamforming at the RRHs, we derive an achievable rate for a randomly selected user in the network. Further, based on the dominant RRH approach, we provide an approximate but accurate expression to analytically evaluate this rate averaged over the spatial realizations of RRH and user locations. From our analysis, we arrive at the following conclusions: (1) the achievable average system sum-rate is a strictly quasi-concave function of the number of users in the network, (2) for the same number of antennas in the system, the optimal number of antennas per RRH to maximize the average user rate as well as average system sum-rate depends on the quality of the CSI. While for a high-quality CSI a more collocated system is preferred, for low-quality CSI it is better to consider a more distributed RRH deployment.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call