点击次数：

论文类型：期刊论文

发表时间：2017-09-30

发表刊物：KSII TRANSACTIONS ON INTERNET AND INFORMATION SYSTEMS

收录刊物：Scopus、SCIE、EI

卷号：11

期号：9

页面范围：4398-4417

ISSN号：1976-7277

关键字：MU-MIMO; user selection; fairness; dynamic time warping

摘要：Multi-user Multiple-Input and Multiple-Output (MU-MIMO) has potential for prominently enhancing the capacity of wireless network by simultaneously transmitting to multiple users. User selection is an unavoidable problem which bottlenecks the gain of MU-MIMO to a great extent. Major state-of-the-art works are focusing on improving network throughput by using Channel State Information (CSI), however, the overhead of CSI feedback becomes unacceptable when the number of users is large. Some work does well in balancing tradeoff between complexity and achievable throughput but is lack of consideration of fairness. Current works universally ignore the rational utilizing of time resources, which may lead the improvements of network throughput to a standstill. In this paper, we propose TOUSE, a scalable and fair user selection scheme for MU-MIMO. The core design is dynamic-timewarping- based user selection mechanism for downlink MU-MIMO, which could make full use of concurrent transmitting time. TOUSE also presents a novel data-rate estimation method without any CSI feedback, providing supports for user selections. Simulation result shows that TOUSE significantly outperforms traditional contention-based user selection schemes in both throughput and fairness in an indoor condition.