Optical nonlinear equalization method based on transfer learning

A technology of transfer learning and equalization method, applied in neural learning methods, electromagnetic wave transmission systems, biological neural network models, etc., can solve problems such as inability to efficiently realize multi-channel, saving data resources and training time, fast modeling, The effect of improving the efficiency of nonlinear equalization

Active Publication Date: 2019-11-26
UNIV OF ELECTRONICS SCI & TECH OF CHINA
View PDF6 Cites 11 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The technical problem to be solved by the present invention is to improve the tolerance of the existing multiplexing optical fiber transmission system to the nonlinear effec

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Optical nonlinear equalization method based on transfer learning
  • Optical nonlinear equalization method based on transfer learning
  • Optical nonlinear equalization method based on transfer learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0021] like figure 2 As shown, the present invention is applied in the digital processing module DSP before each channel decision module after the demultiplexer at the receiver terminal demultiplexes the transmission and reception signals in the same optical fiber into multi-channel signals.

[0022] image 3 Shown is the functional block diagram of the system of the present invention, and concrete steps are as follows:

[0023] 1. In the initialization stage of the system, randomly extract data from each sub-channel to form a set of training data to train the deep neural network model DNNsource, and obtain the initialization neural network parameters;

[0024] 2. Since multiple channels are transmitted in the same optical fiber, the initialization neural network parameters obtained in the initialization stage are migrated to the deep neural network DNNtarget of each channel to assist each channel to quickly establish their own neural network models;

[0025] 3. The deep ne...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention provides an optical nonlinear equalization method based on transfer learning. Trained initialized neural network parameters are migrated to a neural network of each channel through transfer learning, and each channel is assisted to quickly establish the neural network equalizer, so that quick modeling is realized, and the resource overhead, such as less new training data and less iterative compensation, is reduced. Meanwhile, once the channel state is changed, such as the optical power and the transmission distance, the nonlinear phase noise is correspondingly changed, and at themoment, if each channel is retrained respectively, huge expenditure is also corresponding. At the moment, by initializing migration of neural network parameters and supplementing a small amount of new data, rapid neural network remodeling is realized, and then updated parameters are migrated to each channel for updating, so that the response capability to channel change is improved. According tothe invention, nonlinear equalization efficiency under different channels is improved, and high tolerance to optical fiber nonlinearity is maintained.

Description

technical field [0001] The invention relates to a nonlinear equalization technology of a multi-channel optical fiber communication system. Background technique [0002] With the continuous emergence of emerging Internet services such as social networks, cloud computing, and virtual reality, the amount of data generated and the demand for communication capacity are growing exponentially. Government agencies, large enterprises, Internet companies, etc. are actively building their own data centers, and the optical interconnection between data centers and even ultra-large-scale data centers has become a hot research issue. The schemes of direct detection system and coherent detection system using intensity modulation or optical I / Q modulator at the transmitting end emerge in endlessly. Based on single-step linear filtering, Volterra filter, and Kramers-Kronig transceiver schemes, great progress has been made in the compensation of beat frequency noise between signals. The 100G...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): H04B10/61H04B10/69G06N3/08
CPCH04B10/6163H04B10/6971G06N3/08
Inventor 张静夏乐雷平平冯宇中许渤邱昆
Owner UNIV OF ELECTRONICS SCI & TECH OF CHINA
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap