Scholarlink

<p>This paper unveils a cutting-edge end-to-end learning approach to tackle laser phase noise in coherent optical orthogonal frequency-division multiplexing (CO-OFDM) fiber communications. Inspired by the autoencoder concept, a powerful deep neural network, the proposed approach tackles this challenge by learning resilient symbol sequences that can withstand laser phase noise impairments, even from low-cost distributed feedback lasers with linewidths up to 2&#xa0;MHz. This makes it ideally suited for high-order modulation formats and large FFT sizes in CO-OFDM systems, where maximizing spectral efficiency is crucial. By eliminating the need for computationally intensive conventional techniques, this approach opens the door to a simpler, potentially more efficient solution.</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

Leveraging autoencoder for laser phase noise compensation in coherent optical OFDM systems

  • 期刊论文

  • 发表日期：

    2025年1月29日

查看全文
  • Ibtesam R. K. Al-Saedi,
  • Omar Alnaseri,
  • Jamal Mohammed Rasool,
  • Hongxiang Li

摘要

This paper unveils a cutting-edge end-to-end learning approach to tackle laser phase noise in coherent optical orthogonal frequency-division multiplexing (CO-OFDM) fiber communications. Inspired by the autoencoder concept, a powerful deep neural network, the proposed approach tackles this challenge by learning resilient symbol sequences that can withstand laser phase noise impairments, even from low-cost distributed feedback lasers with linewidths up to 2 MHz. This makes it ideally suited for high-order modulation formats and large FFT sizes in CO-OFDM systems, where maximizing spectral efficiency is crucial. By eliminating the need for computationally intensive conventional techniques, this approach opens the door to a simpler, potentially more efficient solution.