Proportional Fair Scheduling for Downlink mmWave Multi-User MISO-NOMA Systems

Autor: Mingshan Zhang, Yongna Guo, Lou Salaun, Chi Wan Sung, Chung Shue Chen
Přispěvatelé: Intel China Research Center, Nokia Bell Labs [Paris-Saclay], City University of Hong Kong [Hong Kong] (CUHK), Laboratory of Information, Network and Communication Sciences (LINCS), Institut National de Recherche en Informatique et en Automatique (Inria)-Institut Mines-Télécom [Paris] (IMT)-Sorbonne Université (SU)
Rok vydání: 2022
Předmět:
Zdroj: IEEE Transactions on Vehicular Technology
IEEE Transactions on Vehicular Technology, 2022, ⟨10.1109/TVT.2022.3159612⟩
ISSN: 1939-9359
0018-9545
DOI: 10.1109/tvt.2022.3159612
Popis: International audience; In this paper, we study non-orthogonal multiple access (NOMA) user scheduling and resource allocation problem for a generic downlink single-cell multiple input and single output (MISO) millimeter wave (mmWave) system. The larger number of packed antennas and the highly directional property of mmWave communications enable directional beamforming to achieve spatial diversity. Toward this end, we consider two different hybrid precoding schemes which are based on orthogonal matching pursuit (OMP). Users are assigned into different clusters and the base station (BS) transmits superposed signals that share the same precoding vector. Moreover, both fixed number of users per cluster and dynamic number of users per cluster are investigated. We aim to jointly optimize the user clustering, service scheduling, and power allocation strategy, in maximizing the proportional fairness (PF) among the users and exploring the multiuser diversity and multiplexing gain. Since the formulated joint user clustering, scheduling and power allocation problem is a mixed integer non-convex optimization problem, we propose a twofold methodology. First, we apply a hybrid precoding and user clustering scheme, where the hybrid precoder is constructed by singular vector division (SVD) or minimum mean square error (MMSE). Then, with the obtained result, we approximate the proportional fairness power allocation problem by a sequence of Geometric Programming (GP) problems which are solved iteratively. The proposed scheme strikes a balance between the spectral efficiency and service fairness. Results show that the proposed MISO-NOMA scheme which is based on MMSE hybrid precoder and the proposed user scheduling and power allocation strategy under proportional fairness metric can outperform various conventional MISO schemes. Furthermore, our proposed dynamic number of users per cluster scheme outperforms the fixed scheme and can be considered as an upper bound in several aspects, including spectral efficiency and fairness.
Databáze: OpenAIRE
Externí odkaz: