Scholarlink

<p>This paper proposes a dual-stage deep learning (DSDL) model for simultaneous transmission parameter identification and optical modulation index (OMI) monitoring for enabling robust radio over fiber (RoF) optical systems. The estimation of OMI at receiving end is essential in calculating the carrier-to-noise ratio (CNR), which is used for transmitter configurations to achieve optimal carrier power for effective transmission. Also, multiparameter monitoring of the optical signals improves the RoF system both in terms of the capacity and helps in increasing the optical transmission distance to deploy efficient RoF systems. In this work, the transmission parameters are identified using the multi-input convolutional neural network (CNN) at stage 1 which is then fed as input for the artificial neural network (ANN) at stage 2 to estimate OMI. Simulation results demonstrate that the DSDL method identifies the transmitting parameters such as the modulation format (MF), center wavelength (CWL), roll-off factor (ROF), transmitting distance (TD), and laser linewidth (LLW) with an average of 99.26% accuracy after training for 200 epochs and the mean absolute error (MAE) of 0.0148 is achieved for OMI estimation. The results obtained are evidence that our proposed approach improves the efficacy of radio over fiber system and its optimization.</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

Dual-Stage Deep Learning Model for Multiparameter Optical Performance Monitoring and Optical Modulation Index Estimation for Radio over Fiber Systems

  • 期刊论文

  • 发表日期：

    2025年1月9日

查看全文
  • Sindhumitha Kulandaivel,
  • R. K. Jeyachitra

摘要

This paper proposes a dual-stage deep learning (DSDL) model for simultaneous transmission parameter identification and optical modulation index (OMI) monitoring for enabling robust radio over fiber (RoF) optical systems. The estimation of OMI at receiving end is essential in calculating the carrier-to-noise ratio (CNR), which is used for transmitter configurations to achieve optimal carrier power for effective transmission. Also, multiparameter monitoring of the optical signals improves the RoF system both in terms of the capacity and helps in increasing the optical transmission distance to deploy efficient RoF systems. In this work, the transmission parameters are identified using the multi-input convolutional neural network (CNN) at stage 1 which is then fed as input for the artificial neural network (ANN) at stage 2 to estimate OMI. Simulation results demonstrate that the DSDL method identifies the transmitting parameters such as the modulation format (MF), center wavelength (CWL), roll-off factor (ROF), transmitting distance (TD), and laser linewidth (LLW) with an average of 99.26% accuracy after training for 200 epochs and the mean absolute error (MAE) of 0.0148 is achieved for OMI estimation. The results obtained are evidence that our proposed approach improves the efficacy of radio over fiber system and its optimization.