High-precision demodulation method for the cobweb FBG sensor network.

Area of Science:

• Optical Engineering
• Signal Processing
• Artificial Intelligence

Background:

• Fiber Bragg Grating (FBG) sensor networks are crucial for precise measurements.
• Demodulating highly overlapping signals in FBG networks presents significant accuracy and speed challenges.
• Existing methods struggle with complex signal interference, limiting network performance.

Purpose of the Study:

• To enhance the demodulation accuracy and speed of cobweb FBG sensor networks.
• To develop a robust algorithm capable of handling highly overlapping sensing signals.
• To reduce signal demodulation errors and shorten detection times.

Main Methods:

• A novel demodulation algorithm integrating improved wavelet adaptive threshold de-noising with a one-dimensional (1D) dilated convolutional neural network (CNN) was developed.
• The wavelet de-noising technique was employed to preprocess highly overlapping optical fiber sensing signals.
• A well-trained 1D dilated CNN model was utilized for the core signal demodulation process.

Main Results:

• The proposed algorithm achieved a root-mean-square error of less than 0.10 pm in determining peak wavelengths for four highly overlapping signals.
• The average demodulation time was significantly reduced to 15.2 ms.
• Experimental results demonstrated substantial improvements in detection accuracy and reduced detection time for the cobweb FBG sensor network.

Conclusions:

• The combined wavelet de-noising and 1D dilated CNN algorithm offers a highly accurate and efficient solution for FBG sensor network demodulation.
• This approach effectively overcomes the challenges posed by highly overlapping signals, enhancing measurement reliability.
• The developed method represents a significant advancement in optical sensing technology, enabling faster and more precise data acquisition.