基于嵌套张量模型的MIMO中继系统组合接收算法

目前在单向双跳多输入多输出（MIMO）中继系统中,基于嵌套张量模型的接收算法主要采用单步交替最小二乘（ALS）和KRF（Khatri-Rao Factorization）算法。在时变信道且实时性要求较高场景下,计算复杂度高是制约其应用的主要因素。为此,在对单向双跳MIMO中继系统建模基础上,提出了基于嵌套张量模型的双步组合接收算法。该算法通过对接收的数据张量进行重建,将符号估计和信道估计分离,充分利用ALS和KRF的算法优势,有效降低了计算复杂度。同时,对算法的可辨识性进行了分析。仿真结果表明,该算法保持了与传统嵌套PARAFAC的最小二乘（Nested PARAFAC ALS）算法的相同估计...

Full description

Saved in:
Bibliographic Details
Published in电讯技术 Vol. 57; no. 8; pp. 885 - 891
Main Author 王瑞 芮国胜 张洋
Format Journal Article
LanguageChinese
Published 海军航空工程学院 电子信息工程系,山东 烟台,264001 2017
Subjects
中继系统
半盲接收
多输入多输出
嵌套张量模型
多输入多输出
嵌套张量模型
semi-blind receiver
multiple-input multiple-output ( MIMO )
nested tensor model
中继系统
relay system
半盲接收
Online AccessGet full text

Cover

More Information
Summary:目前在单向双跳多输入多输出（MIMO）中继系统中,基于嵌套张量模型的接收算法主要采用单步交替最小二乘（ALS）和KRF（Khatri-Rao Factorization）算法。在时变信道且实时性要求较高场景下,计算复杂度高是制约其应用的主要因素。为此,在对单向双跳MIMO中继系统建模基础上,提出了基于嵌套张量模型的双步组合接收算法。该算法通过对接收的数据张量进行重建,将符号估计和信道估计分离,充分利用ALS和KRF的算法优势,有效降低了计算复杂度。同时,对算法的可辨识性进行了分析。仿真结果表明,该算法保持了与传统嵌套PARAFAC的最小二乘（Nested PARAFAC ALS）算法的相同估计性能,在源天线个数变化时,计算复杂度降低了80％以上;在中继天线个数变化时,计算复杂度降低了50％以上。
Bibliography:relay system;semi-blind receiver;nested tensor model
The alternating least squares（ALS） algorithm and Khatri-Rao factorization（KRF） algorithm, which are one-step algorithms,are mainly adopted by receiver algorithms based on nested tensor in multi-ple-input multiple-output（ MIMO） relay communication systems. The high computational cost is a domi-nant factor restricting their application under time-varying channel and real-time conditions. For this rea-son,an assembled receiver algorithm based on nested tensor model is proposed by building one-way and two-hop MIMO relay communication system. The processes of symbol estimation and channel estimation are separated by reconstructing received signal,and the computation complexity is reduced effectively. Mo-reover,identifiability conditions of the algorithm are analyzed. Simulation results show that the computation complexity of the algorithm is decreased more than 80% and 50% with different source antennas and relay antennas,while the algorithm has the same estimati
ISSN:1001-893X
DOI:10.3969/j.issn.1001-893x.2017.08.006