MIMO detection using a deep learning neural network in a mode division multiplexing optical transmission system

被引：19

作者：

Poudel, Bishal ^{[1
]}

Oshima, Joji ^{[1
]}

Kobayashi, Hirokazu ^{[1
]}

Iwashita, Katsushi ^{[1
]}

机构：

[1] Kochi Univ Technol, Dept Photon Syst Engn, 185 Miyanokuchi, Kami City, Kochi 7828502, Japan

来源：

OPTICS COMMUNICATIONS | 2019年 / 440卷

关键词：

Multiple Input Multiple Output; Deep learning neural network; Mode division multiplexing; Optical transmission system; Optimum detector; SDM TRANSMISSION; EQUALIZATION; SIGNALS; CORE;

D O I：

10.1016/j.optcom.2019.02.016

中图分类号：

O43 [光学];

学科分类号：

070207 ; 0803 ;

摘要：

Optimum Multiple Input Multiple Output (MIMO) detector has always been a challenge in MIMO communication systems. In this paper, a novel MIMO detector has been designed using a supervised Deep Learning Neural Network (DLNN) and has been implemented successfully in a Mode Division Multiplexing (MDM) optical transmission system. A conventional Graded-Index Multi-Mode Fiber (GI-MMF) is used to design an MDM optical transmission system. We have used a DLNN for MIMO detection in MDM optical transmission system and have compared its performance with Zero Forcing (ZF) detector and Semi-Definite Relaxation Row-by-Row (SDR-RBR). The results confirm that our DLNN outruns the performance of traditional MIMO detectors.

引用

页码：41 / 48

页数：8

共 29 条

[1]

Abadi M, 2015, TENSORFLOW LARGE SCA

[2] LP01/LP11 dual-mode and dual-polarisation CO-OFDM transmission on two-mode fibre [J].