Related Experiment Video
Updated: Jun 29, 2025
11:30
Recombination Dynamics in Thin-film Photovoltaic Materials via Time-resolved Microwave Conductivity
Published on: March 6, 2017
11.7K
Microwave photonics frequency measurement with improved accuracy based on an artificial neural network
Applied Optics
|April 3, 2024
Summary
This study introduces an artificial neural network (ANN) to enhance photonics-assisted microwave frequency measurement (MFM) accuracy. The ANN method significantly reduces errors and improves frequency resolution in multi-tone measurements.
Area of Science:
- Photonics
- Microwave Engineering
- Artificial Intelligence
Background:
- Photonics-assisted microwave frequency measurement (MFM) offers solutions to electronic limitations in radar and communication.
- Stimulated Brillouin scattering (SBS) based MFM systems can measure multiple high-frequency signals but face accuracy constraints due to SBS bandwidth and system nonlinearity.
Purpose of the Study:
- To improve the accuracy and resolution of multi-tone frequency measurements in SBS-based MFM systems.
- To address the limitations imposed by SBS bandwidth and system nonlinearity in high-frequency signal analysis.
Main Methods:
- An artificial neural network (ANN) was developed to establish a nonlinear mapping between signal spectra and theoretical frequencies.
- The ANN was applied to optimize frequency measurements within the 0.5–27 GHz range for MFM systems.
Main Results:
- The ANN method demonstrated significant error reductions across various signal-to-noise ratio (SNR) conditions: 79% (20 dB), 76% (15 dB), 70% (10 dB), and 44% (0 dB).
- Frequency resolution was enhanced from 30 MHz to 10 MHz.
Conclusions:
- ANN-based optimization effectively overcomes nonlinearities and bandwidth limitations in SBS-based MFM.
- This approach substantially enhances the precision and resolution of microwave frequency measurements, particularly for multi-tone signals.
