Multimode Fiber Specklegram Sensor for Multi-Position Loads Recognition Using Traversal Occlusion.

Area of Science:

• Photonics and Optical Sensing
• Machine Learning for Sensor Systems
• Distributed Fiber Optic Sensing

Background:

• Simultaneous measurement of multiple perturbation positions and intensities in multimode fiber (MMF) distributed sensors is challenging due to large specklegram data requirements.
• Existing methods face difficulties in recognizing multiple perturbations efficiently, hindering the development of high-resolution distributed sensing applications.

Purpose of the Study:

• To propose and validate a novel approach for recognizing multi-position loads using MMF specklegram sensors.
• To enhance the sample diversity and robustness of MMF-distributed sensor recognition models.
• To provide a cost-effective and efficient solution for high-resolution distributed measurements.

Main Methods:

• Development of a multi-variable, multi-class, one-shot specklegram dataset construction method.
• Theoretical derivation of the mathematical model for total local intensity and its sensitivity to perturbations.
• Implementation of specklegram traversal occlusion data augmentation with a shallow convolutional neural network (CNN).

Main Results:

• The proposed method achieves nearly 100% accuracy in simultaneously recognizing load positions and magnitudes for up to 1545 distinct load forms.
• The shallow CNN model demonstrates superior training efficiency and stability compared to existing MMF sensing models.
• Experimental validation confirms the effectiveness of the approach for distributed sensing applications.

Conclusions:

• The study presents a proof of concept for a distributed MMF specklegram sensor capable of high-resolution measurements under diverse perturbations.
• The developed method significantly advances MMF-based distributed sensing by offering enhanced accuracy and efficiency.
• This approach provides a promising, cost-effective solution for various distributed sensing applications.