Novel Phase-Noise-Tolerant Variational-Autoencoder-Based Equalization Suitable for Space-Division-Multiplexed Transmission

2025-09-17

AI Summary

In spatial-division-multiplexing (SDM) systems employing randomly coupled multi-core fiber, severe phase noise severely degrades performance, while conventional equalization methods suffer from insufficient modeling capability. To address this, we propose a novel nonlinear equalization scheme based on variational autoencoders (VAEs), the first to apply VAEs to phase-noise-tolerant SDM equalization. By leveraging latent-variable modeling, the method enables adaptive representation and compensation of nonlinear phase distortions induced by the channel. Unlike linear or deterministic deep-learning equalizers, the VAE effectively captures both the statistical characteristics of phase noise and the dynamic spatial coupling among cores. In a 150-km randomly coupled multi-core fiber experiment, the proposed scheme improves optical signal-to-noise ratio (OSNR) tolerance by 2.3 dB over conventional MIMO-LMS and CNN-based equalizers, significantly enhancing transmission stability and phase robustness. This work establishes a scalable, intelligent equalization paradigm for high-dimensional SDM systems.

Technology Category

• Machine Learning: Deep Generative Models & Autoencoders
• Computer Vision: Learning & Optimization for CV
• Search and Optimization: Non-convex Optimization

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Systems and Infrastructure for Web, Mobile and WoT: Virtualization and resource management in Web systems and infrastructures

Abstract

We demonstrate the effectiveness of a novel phase-noise-tolerant, variational-autoencoder-based equalization scheme for space-division-multiplexed (SDM) transmission in an experiment over 150km of randomly-coupled multi-core fibers.

Problem

• Equalization for space-division-multiplexed transmission
• Tolerant to phase noise in optical fibers
• Using variational autoencoder-based compensation scheme

Innovation

• Phase-noise-tolerant variational autoencoder equalization
• Space-division-multiplexed transmission enhancement
• Randomly-coupled multi-core fiber experiment