Scholarlink

<p>In recent times, there has been a huge amount of work on utilizing deep learning (DL) to estimate the quality of transmission (QoT) in optical networks. This research depict a lightpath’s quality of transmission to develop advanced fiber-optic communication and networks based on DL technique. We need different primary estimation parameters for advanced optical fiber communication and networks, i.e., modulation formats, baud rate, and code rate.Recently, the QoT for unspecified optical paths relies on various estimation approaches i.e., (1) analytical models estimating physical layer impairments (PLIs) and (2) margined formulas. This paper emphasis on Gaussian mixture model (GMM) based algorithm that can be applied to optimization and sophisticated systems. The model can forecast bit-error rate, and signal-to-noise ratio (SNR) of unknown optical paths with threshold value, traffic volume, and modulation format. The model was trained and tested using features from Korean network topology. The Area Under the ROC Curve (AUC) from the simulated outcome is 1.00, while maintaining the high accuracy, F1 score, low Brier score and low expected calibration error (ECE).</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Gaussian mixture model for enhancing the quality of transmission estimation in optical networks: a machine learning approach

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年12月9日

查看全文
  • Shakrajit Sahu,
  • J. Christopher Clement

摘要

In recent times, there has been a huge amount of work on utilizing deep learning (DL) to estimate the quality of transmission (QoT) in optical networks. This research depict a lightpath’s quality of transmission to develop advanced fiber-optic communication and networks based on DL technique. We need different primary estimation parameters for advanced optical fiber communication and networks, i.e., modulation formats, baud rate, and code rate.Recently, the QoT for unspecified optical paths relies on various estimation approaches i.e., (1) analytical models estimating physical layer impairments (PLIs) and (2) margined formulas. This paper emphasis on Gaussian mixture model (GMM) based algorithm that can be applied to optimization and sophisticated systems. The model can forecast bit-error rate, and signal-to-noise ratio (SNR) of unknown optical paths with threshold value, traffic volume, and modulation format. The model was trained and tested using features from Korean network topology. The Area Under the ROC Curve (AUC) from the simulated outcome is 1.00, while maintaining the high accuracy, F1 score, low Brier score and low expected calibration error (ECE).