Related Experiment Video
Updated: Jun 24, 2025
16:11
Implementation of a Reference Interferometer for Nanodetection
Published on: April 26, 2014
9.4K
Reflective microring-resonator-based microwave photonic sensor incorporating a self-attention assisted convolutional
Applied Optics
|June 10, 2024
Summary
A novel microwave photonic sensor using a reflective microring resonator and a self-attention CNN achieves a 10x improvement in temperature sensing accuracy. This advanced sensor leverages optical resonant modes for precise measurements, even with limited data.
Area of Science:
- Photonics
- Optical Sensing
- Machine Learning
Background:
- Microring resonators (MRRs) are key components in photonic integrated circuits.
- Microwave photonic (MWP) sensors offer high sensitivity for various measurements.
- Convolutional neural networks (CNNs) excel at analyzing spectral data.
Purpose of the Study:
- To develop a highly accurate MWP sensor using an MRR probe.
- To integrate a self-attention CNN for enhanced signal processing.
- To demonstrate improved sensing performance, particularly in temperature sensing applications.
Main Methods:
- Utilizing an MRR cascaded with an inverse-designed optical reflector as the sensor probe.
- Employing MWP interrogation to convert resonant modes into radio-frequency (RF) spectral notches.
- Applying a self-attention CNN to analyze RF spectral features for accurate sensing.
Main Results:
- The proposed sensor successfully converted resonant modes into distinctive RF spectral notches.
- The self-attention CNN model achieved a root-mean-square error of 0.026°C in temperature sensing.
- A 10-fold improvement in sensing accuracy was observed compared to traditional linear fitting models.
Conclusions:
- The developed MRR-based MWP sensor with self-attention CNN demonstrates superior accuracy.
- This approach effectively leverages both local and global features of RF spectra.
- The method shows significant potential for high-precision sensing applications, even with small datasets.
