Scholarlink

Overcoming the fiber nonlinearity is the most significant challenge, which constraints the quality of the optical fiber communication schemes. Nevertheless, the transmission signals are the focus to significant interface because of the mode coupling as well as dispersion, which need the efficient digital signal processing approaches for restoring the quality of the transmission signals. Hence, this research proposes the Multilayer Feedforward Neural Network (MLFNN) method for the adaptive nonlinear equalization for the fiber optic channels. MLFNN learn directly from data, minimizing the requirement for comprehensive analytical modeling of the nonlinear channel. This makes them adjustable to various systems as well as operational conditions. The efficacy of the proposed MLFNN method is estimated with Peak Signal Noise Ratio (PSNR). The proposed MLFNN approach attains the better PSNR of 29.7 dB and 28.8 dB on the transmission distances of 600 and 800 km, respectively, when compared to the existing methods like Perturbation-based Neural Network (PNN).

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

Adaptive Non-linear Equalization for the Fiber-Optic Channels Using Multilayer Feedforward Neural Network

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Mohammed Al-Farouni,
  • S. Abdul Ameer,
  • Deepthi Palakurthy

摘要

Overcoming the fiber nonlinearity is the most significant challenge, which constraints the quality of the optical fiber communication schemes. Nevertheless, the transmission signals are the focus to significant interface because of the mode coupling as well as dispersion, which need the efficient digital signal processing approaches for restoring the quality of the transmission signals. Hence, this research proposes the Multilayer Feedforward Neural Network (MLFNN) method for the adaptive nonlinear equalization for the fiber optic channels. MLFNN learn directly from data, minimizing the requirement for comprehensive analytical modeling of the nonlinear channel. This makes them adjustable to various systems as well as operational conditions. The efficacy of the proposed MLFNN method is estimated with Peak Signal Noise Ratio (PSNR). The proposed MLFNN approach attains the better PSNR of 29.7 dB and 28.8 dB on the transmission distances of 600 and 800 km, respectively, when compared to the existing methods like Perturbation-based Neural Network (PNN).