Estimation of variable optical feedback coupling factor for self-mixing interferometry by signal-to-image translation

Area of Science:

• Optics and Photonics
• Machine Learning
• Sensor Technology

Background:

• Accurate estimation of the optical feedback coupling factor (C) is critical for self-mixing interferometry (SMI) sensor reliability.
• Variable feedback conditions pose challenges for traditional SMI parameter estimation methods.

Purpose of the Study:

• To develop a novel deep learning-based method for estimating the time-varying optical feedback coupling factor (C) in SMI sensors.
• To enhance the robustness and accuracy of C factor estimation under diverse feedback regimes (strong, moderate, weak).

Main Methods:

• Transforming one-dimensional (1D) SMI signals into two-dimensional (2D) image representations.
• Utilizing convolutional neural networks (CNNs) for feature extraction from the 2D signal images.
• Formulating the estimation as a signal-to-image translation task for sensor parameter inference.

Main Results:

• The proposed 2D deep learning method demonstrates superior performance compared to state-of-the-art recurrent neural network models (LSTM, Transformer).
• The approach exhibits enhanced robustness against variations in sampling frequency, displacement amplitude, and feedback regime.
• Experimental validation confirms the effectiveness of the 2D-based signal-to-image translation for C factor estimation.

Conclusions:

• The novel 2D deep learning method offers a robust and accurate solution for estimating the optical feedback coupling factor in SMI sensors.
• This methodology is generalizable and applicable to various sensor signal analysis tasks.
• A publicly available implementation facilitates reproducibility and practical adoption of the approach.