Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

PIDT: Physics-Informed Digital Twin for Optical Fiber Parameter Estimation

Zicong Jiang, Magnus Karlsson, Erik Agrell, Christian Häger

TL;DR

The paper tackles estimating fiber parameters $\alpha$, $\beta_2$, and $\gamma$ along an optical link using realistic signals. It proposes PIDT, a physics-informed digital twin that embeds a parameterized SSFM as a differentiable twin within a neural-operator framework and uses natural bicubic splines to interpolate discrete SSFM outputs to continuous $(z,t)$ coordinates while enforcing the NLSE residual via a physics-informed loss. The approach delivers higher estimation accuracy and faster convergence with lower parameter count than PINO, and yields physically-consistent parameter profiles along the fiber compared with loss-only SSFM baselines. This enables efficient impairment-aware fiber parameter estimation for monitoring, compensation, and optimization in optical communications.

Abstract

We propose physics-informed digital twin (PIDT): a fiber parameter estimation approach that combines a parameterized split-step method with a physics-informed loss. PIDT improves accuracy and convergence speed with lower complexity compared to previous neural operators.

PIDT: Physics-Informed Digital Twin for Optical Fiber Parameter Estimation

TL;DR

The paper tackles estimating fiber parameters , , and along an optical link using realistic signals. It proposes PIDT, a physics-informed digital twin that embeds a parameterized SSFM as a differentiable twin within a neural-operator framework and uses natural bicubic splines to interpolate discrete SSFM outputs to continuous coordinates while enforcing the NLSE residual via a physics-informed loss. The approach delivers higher estimation accuracy and faster convergence with lower parameter count than PINO, and yields physically-consistent parameter profiles along the fiber compared with loss-only SSFM baselines. This enables efficient impairment-aware fiber parameter estimation for monitoring, compensation, and optimization in optical communications.

Abstract

We propose physics-informed digital twin (PIDT): a fiber parameter estimation approach that combines a parameterized split-step method with a physics-informed loss. PIDT improves accuracy and convergence speed with lower complexity compared to previous neural operators.
Paper Structure (4 sections, 1 equation, 2 figures)

This paper contains 4 sections, 1 equation, 2 figures.

Table of Contents

  1. Introduction
  2. Background and related approaches
  3. Proposed physics-informed digital twin (PIDT) for fiber parameter estimation
  4. Numerical Results

Figures (2)

  • Figure 1: Fiber-parameter estimation using (a) the parameterized SSFM, (b) PINO, and (c) PIDT (proposed).
  • Figure 2: Estimation results comparing PIDT-$N_\text{sym}$$(M=4)$ with (a) PINO-$N_\text{sym}$ for $\hat{\theta}$ and (b) parameterized SSFM for $\theta_\text{DT}$. Both $\text{PINO}_{\text{small}}$ ($3$ layers, $200$ neurons/layer) and $\text{PINO}_{\text{large}}$ ($7$ layers, $900$ neurons/layer) use tanh activation.