Network-Assisted Full-Duplex Millimeter-Wave Cell-Free Massive MIMO with Localization-Aided Inter-User Channel Estimation

Abstract

We propose a joint resource allocation and beam-forming design with location-aided channel estimation to mitigate the inter-user interference in network-assisted full-duplex (NAFD) cell-free (CF) massive multiple-input multiple-output (mMIMO) systems operating in millimeter-wave (mmWave) channels. The key idea is to utilize the approximate estimates of the channels between users that can be obtained from knowledge of user locations, since mmWave channels are dominated by line-of-sight (LoS) paths due to their high propagation loss nature. This rough channel state information (CSI) enables the system to judiciously choose access points (APs) and design beamformers to significantly enhance the performance of the NAFD CF-mMIMO system. Simulation results confirm that the total throughput of the proposed methods is close to that with the perfect channel knowledge between inter-user channels.