Nonlinear equalization method based on gated recurrent neural network

Abstract

The invention discloses a nonlinear equalization method based on a gated recurrent neural network. The method comprises the steps of determining a gated recurrent neural network model, training the gated recurrent neural network model by using the training sample data, and obtaining a gated recurrent neural network model after preliminary training; performing optimization processing on the gatingrecurrent neural network model generated by the preliminary training; trimming the gated recurrent neural network model after optimization processing; retraining the trimmed gated recurrent neural network model by using the training sample data to obtain a trained gated recurrent neural network model; and carrying out nonlinear estimation or equalization processing by using the trained gated recurrent neural network model, thereby realizing a high-performance and low-complexity nonlinear loss compensation function.

Description

technical field [0001] The invention relates to the technical field of optical communication, in particular to a nonlinear equalization method based on a gated recurrent neural network. Background technique [0002] The nonlinear effects existing in the optical transmission system and optical access network have a great impact on signal transmission performance and signal quality. The traditional nonlinear equalization algorithm is very complex, depends on link information, is difficult to implement, and has great limitations in application. Although the existing nonlinear equalization algorithms based on machine learning can achieve similar performance to traditional algorithms with lower complexity, most of the nonlinear equalization algorithms based on artificial neural networks and support vector machines have large training costs and poor convergence. Due to the problems of slow speed and small improvement of Q factor, the clustering scheme based on fuzzy logic C-means...

AI Technical Summary

Problems solved by technology

The traditional nonlinear equalization algorithm is very complex, depends on link information, is difficult to implement, and has great limitations in application

Although the existing nonlinear equalization algorithms based on machine learning can achieve similar performance to traditional algorithms with lower complexity, most of the nonlinear equalization algorithms based on artificial neural networks and support vector machines have large training costs and poor convergence. The problem of slow speed and small improvement of Q factor, the clustering scheme based on fuzzy logic C-means has great limitations in the order of modulation format used in system transmission

Images