计算机工程 ›› 2019, Vol. 45 ›› Issue (12): 58-63.doi: 10.19678/j.issn.1000-3428.0055031

• 移动互联与通信技术 • 上一篇    下一篇

多小区分布式阵列系统训练模式与协作方法研究

张帅, 王璀, 潘鹏, 严军荣   

  1. 杭州电子科技大学 通信工程学院, 杭州 310018
  • 收稿日期:2019-05-27 修回日期:2019-07-20 发布日期:2019-08-09
  • 作者简介:张帅(1995-),男,硕士研究生,主研方向为MIMO预编码技术;王璀,硕士研究生;潘鹏,副教授;严军荣,讲师。
  • 基金项目:
    国家自然科学基金(U1709220)。

Research on Training Mode and Collaborative Method of Multi-Cell Distributed Array Systems

ZHANG Shuai, WANG Cui, PAN Peng, YAN Junrong   

  1. School of Communication Engineering, Hangzhou Dianzi University, Hangzhou 310018, China
  • Received:2019-05-27 Revised:2019-07-20 Published:2019-08-09

摘要: 在多小区分布式阵列系统中,用户与接入节点(AP)间的距离对频谱效率具有重要影响,边缘用户可以通过不同小区内AP间的协作传输来获得较高的频谱效率。为此,将多小区的边缘用户以及AP组成一个虚拟簇,针对频分双工模式下簇内多用户信道的训练模式与多AP的协作预编码问题,提出一种基于导频重用的信道训练方法。当AP配置较多的接入天线时,服务于同一个用户的多个AP可以重用同一个导频序列,同时结合相应的预编码来抑制用户间的干扰。仿真结果表明,与正交训练完全协作、正交训练部分协作2种方法相比,该方法能够节约导频资源,降低反馈开销与计算复杂度,在对训练和反馈开销进行加权后,其频谱效率可以得到有效提升。

关键词: 分布式阵列, 边缘用户, 训练模式, 导频重用, 预编码, 有限反馈

Abstract: In the multi-cell distributed array systems,the distance between users and Access Point(AP) has a significant impact on spectrum efficiency.Therefore,edge users can achieve cooperative transmission through AP in different cells to obtain high spectrum efficiency.To this end,this paper combines edge users and AP together as a virtual cluster.To address the problem of multi-user channel training mode and multiple AP cooperative precoding in Frequency Division Duplex(FDD) mode,this paper proposes a channel training method based on pilot reuse.In this method,when AP is configured with lots of access antennas,the multiple AP which serves the same user can reuse the same pilot sequence.At the same time,the corresponding precoding can be combined with AP to suppress the interference between users.Simulation results show that compared with the orthogonal training full coordination and partial coordination methods,the proposed method can save pilot resources,reduce the feedback overhead and computational complexity,and greatly improve the spectrum efficiency after weighting the training and feedback overhead.

Key words: distributed array, edge user, training mode, pilot reuse, precoding, limited feedback

中图分类号: