Association in Dense Cell-Free mmWave Networks

We exploit a dense cell-free mmWave network where User Equipments (UEs) are served by multiple highly directional beams provided by multiple Base Stations (BSs) simultaneously. Such multi-beam scenarios can either offer high spectral efficiency when different information is transmitted through each beam or a diversity gain when each beam transmits the same information. However, this increased spectral efficiency or diversity gain costs a more complex network association phase. A UE requires finding multiple nearby serving BSs and determining the optimal beam pair for each one. Thus, an efficient association process is urgently needed. In this work, we propose a UE-initiated association method for dense cell-free mmWave networks. We design an efficient beam training mechanism with multiple BSs using hybrid beamforming. We evaluate the proposed association method under different network configurations. The simulation results show that compared to traditional solutions, our proposed association method can lead to maximally 100% faster beam training and reduce energy consumption by up to 77%. The proposed UE-initiated association method is also scalable to the number of RF chains and antennas at BSs and UEs, making it very suitable for dense cell-free networks.