Access Point Location Design in Cell-Free Massive MIMO Systems

Abstract

Cell-free massive multiple-input multiple-output (MIMO) is a special deployment of massive MIMO systems with a large number of distributed single antenna access points (APs) serving a much smaller number of users, whose performance is largely influenced by the AP locations. We formulate two AP location optimization problems to maximize the sum-throughput and the minimum-throughput in uplink transmission of cell-free systems with an arbitrary user distribution using zero-forcing detector. Numerical results show that the proposed algorithm for the sum-throughput criterion provides significantly higher sum-throughput over a cell-free system with randomly located users. However, the simple Lloyd algorithm can provide higher 95%-likely per-user rates. An AP location design algorithm is also presented in an alternative small-cell system in which each AP serves only one user. The proposed algorithm is based on the k-means clustering algorithm with a cost function defined according to the average interference from other users in addition to the distances between APs and the served users. The algorithm improves the 95%-likely and minimum rates over the Lloyd algorithm, while the latter remains an attractive approach due to its simplicity and good performance.