Related Experiment Video
Updated: Aug 8, 2025
09:10
Construction and Characterization of External Cavity Diode Lasers for Atomic Physics
Published on: April 24, 2014
27.8K
Forecasting the chaotic dynamics of external cavity semiconductor lasers
Optics Letters
|March 1, 2023
Summary
We developed a photonic reservoir computing method for predicting chaotic time series from semiconductor lasers. This approach uses a single laser node and achieves accurate predictions for over 2 nanoseconds.
Area of Science:
- Nonlinear Dynamics and Photonics
- Complex Systems and Machine Learning
Background:
- Chaotic time series prediction is crucial for numerous applications.
- Existing methods often require complex setups or extensive data.
Purpose of the Study:
- To present a novel single-node photonic reservoir computing approach for chaotic time series forecasting.
- To investigate the prediction performance of chaotic behavior in external cavity semiconductor lasers.
Main Methods:
- Employed a semiconductor laser with optical feedback as the sole nonlinear node in a photonic reservoir.
- Investigated the impact of reservoir meta-parameters on prediction accuracy.
- Utilized only observed data for forecasting.
Main Results:
- Identified an optimal meta-parameter space for forecasting optical-feedback-induced chaos.
- Achieved continuous prediction of chaotic time series for over 2 nanoseconds.
- Obtained a normalized mean squared error below 0.1.
Conclusions:
- The proposed single-node photonic reservoir computing offers a hardware-friendly and effective method for complex chaos prediction.
- This work provides a roadmap for meta-parameter selection in delay-based photonic reservoirs for optimal performance.
