大绝热DML驱动的56 Gbit/s PAM-4信号高效分段线性均衡器：ECOC 2019进展
167 浏览量
更新于2024-08-26
收藏 580KB PDF 举报
本文主要探讨了"使用具有大绝热using的DML进行56 Gbit/s PAM-4信号传输的分段线性均衡器"的研究成果，发表在光通信领域的国际会议ECOC 2019上。作者团队由严复、孔德明、辛海云、贾佳等人组成，分别来自上海交通大学的先进光学通信系统与网络实验室、丹麦技术大学的DTU Fotonik以及杭州电子科技大学通信工程学院。
该研究的核心内容是针对直接调制激光(DML)系统中的一个挑战——由于DML的啁啾(chirp)和色散导致的幅度依赖偏差。为了提高信号传输的性能并保持高效计算，研究人员首次提出了一个计算效率高的分段线性均衡器(Piecewise Linear Equalizer, PLE)设计。通过采用大绝热using技术，能够在C波段实现56 Gbit/s的PAM-4信号传输，这在一定程度上克服了传统系统在长距离（如40公里和60公里）传输时的限制。
DML-DD（Direct Modulation Laser with Direct Detection）因其低成本特性而备受关注，特别是在短距离光纤通信中。然而，随着数据速率的提升，如56 Gbit/s，对信号质量的要求更为严格，因此对线性均衡器的设计和优化至关重要。文章中提到的分段线性均衡器通过将复杂的线性补偿分解成多个简单部分，有效降低了计算复杂度，同时提高了信号的均衡效果，使得在较长距离下也能维持高质量的信号传输。
此外，文中还提到了关键的关键词，如"线性均衡器"、"直接调制激光器"、"脉冲幅度调制"、"直接检测"以及"大绝热啁啾"，这些都揭示了研究的核心技术点。这项工作不仅推动了光通信领域的大数据速率传输技术的发展，也为解决未来高速光纤通信系统中的信号处理问题提供了新的解决方案。
总结来说，本文的创新之处在于提出了一种高效的分段线性均衡器，解决了DML系统中信号传输过程中的特定问题，对于推动光通信技术向更高数据速率和更远距离发展具有重要意义。
1
PIECEWISE LINEAR EQUALIZER FOR 56 GBIT/S
PAM-4 SIGNAL TRANSMISSION USING DML WITH
LARGE ADIABATIC CHIRP
Yan Fu
1,2
, Deming Kong
2
, Haiyun Xin
1,2
, Shi Jia
2
, Kuo Zhang
1
, Meihua Bi
3
,
Weisheng Hu
1
, Hao Hu
2*
1
State Key lab of Advanced Optical Communication Systems and Networks,
Shanghai Jiao Tong University, Shanghai, China
2
DTU Fotonik, Technical University of Denmark, Kgs. Lyngby, Denmark
3
School of Communication Engineering, Hangzhou Dianzi University, Hangzhou, Zhejiang, China
*huhao@fotonik.dtu.dk
Keywords: LINEAR EQUALIZER, DIRECTLY-MODULATED LASER, PULSE-AMPLITUDE
MODULATION, DIRECT DETECTION, ADIABATIC CHIRP
Abstract
Computationally efficient piecewise linear (PWL) equalizer is, for the first time, proposed to compensate amplitude-dependent
skews caused by DML chirp and chromatic dispersion. 56 Gbit/s PAM-4 signals in C-band are respectively enabled for 40 km
and 60 km transmission below the HD-FEC and SD-FEC.
1 Introduction
Pulse-amplitude modulation (PAM) based directly-modulated
laser and direct detection (DML-DD) system has been widely
studied for short reach optical communications due to its low
cost and simple configuration [1]. However, the directly-
modulated PAM signals can easily suffer from distortions after
transmission. Affected by the adiabatic chirp from the DML
working in a high output power region, different intensity
levels of the PAM signals are frequency shifted respectively,
thus resulting in a tilted eye diagram (i.e., amplitude-
dependent skews) due to chromatic dispersion from the
transmission fiber. Various solutions have been proposed to
address this issue, including Volterra series based nonlinear
equalizers [2] and neural network based equalizers [3].
However, the Volterra equalizer requires high-order
polynomials to approximate the nonlinear transmission
response and is computationally inefficient. Neural network-
based equalizer requires intense training with large amounts of
computing resources. Alternative solutions with lower
computational complexity incorporate with two steps
including skew estimation and skew compensation [4-5].
These solutions are based on amplitude classification, which
are not accurate enough for the nonlinear response with a small
number of classification thresholds.
We propose, for the first time, a computational efficient
piecewise linear (PWL) equalizer to rectify the amplitude-
dependent skews due to the interaction between DML chirp
and fiber chromatic dispersion. Enabled by the PWL equalizer,
a 56 Gbit/s PAM-4 signal using DML at high output power
region with an extinction ratio (ER) of 5.18 dB is transmitted
over 40 km below the hard-decision (HD-FEC) threshold; and
60 km below the soft-decision forward error correction (SD-
FEC) threshold.
2 Principle of PWL equalizer
The PWL equalizer is originally proposed for acoustic echo
cancellation where loudspeaker nonlinearities result in a
nonlinear acoustic echo that distorts the output signal [6]. The
equalizer performs well for memoryless nonlinearities, where
adiabatic chirp from DML applies [7]. The PWL equalizer
consists of three steps, including amplitude threshold
decomposition, linear multichannel equalization and linear
addition.
Fig. 1 Workflow of the PWL equalizer.
For the amplitude threshold decomposition, a received
symbol can be represented as a vector of segments defined by
a set of pre-defined thresholds. Fig. 1 shows an example of a
signal sequence that
has been segmented using a threshold set .
The signal is therefore partitioned for four intervals of
(-
∞,
-
1.5), [-1.5, 0), [0, 1.5) and [1.5, ∞). The third element
(2.3) in the sequence is split into four segments as
, where the values are obtained as
(the subtraction of the value and the
threshold),
,
, and
.
In general, continuous threshold decomposition is an operation
that maps real numbers into vectors using a set of predefined
thresholds values. Considering , and the strictly ordered
set of N thresholds
. (1)
下载后可阅读完整内容，剩余3页未读，立即下载
2021-03-06 上传
2021-02-25 上传
2021-05-12 上传
2021-03-07 上传
2021-02-23 上传
106 浏览量
2021-05-13 上传
2021-03-15 上传
2010-12-12 上传
weixin_38635449
- 粉丝: 5
- 资源: 971
我的内容管理
展开
最新资源
- WordPress作为新闻管理面板的实现指南
- NPC_Generator：使用Ruby打造的游戏角色生成器
- MATLAB实现变邻域搜索算法源码解析
- 探索C++并行编程：使用INTEL TBB的项目实践
- 玫枫跟打器：网页版五笔打字工具，提升macOS打字效率
- 萨尔塔·阿萨尔·希塔斯：SATINDER项目解析
- 掌握变邻域搜索算法：MATLAB代码实践
- saaraansh: 简化法律文档，打破语言障碍的智能应用
- 探索牛角交友盲盒系统：PHP开源交友平台的新选择
- 探索Nullfactory-SSRSExtensions: 强化SQL Server报告服务
- Lotide：一套JavaScript实用工具库的深度解析
- 利用Aurelia 2脚手架搭建新项目的快速指南
- 变邻域搜索算法Matlab实现教程
- 实战指南：构建高效ES+Redis+MySQL架构解决方案
- GitHub Pages入门模板快速启动指南
- NeonClock遗产版：包名更迭与应用更新
