This study aims to address the challenge of accurately separating partially and fully overlapped spectra in fiber Bragg grating (FBG) sensors, which is crucial for precise measurements in densely multiplexed sensor networks.
The proposed approach leverages the nonlinear least squares curve fitting method, implemented with the Levenberg–Marquardt algorithm (LMA), to detect distinct FBG peaks. Two FBG sensors were used in the investigation, with one sensor exposed to varying temperatures while the other was kept at a constant temperature as a reference. Experimental measurements were performed across six temperature conditions to evaluate the algorithm's performance in detecting overlapped peaks.
The proposed algorithm demonstrated exceptional accuracy in detecting FBG peaks, even in complex spectral superpositions. The experimental results showed that partially and fully overlapped spectra could be effectively resolved, with measurements achieving a high level of precision across various temperature conditions.
This study presents a novel application of the LMA-based curve fitting method for peak detection in overlapped FBG spectra. Unlike machine learning approaches, the method is computationally efficient, offering a rapid and accurate solution for applications requiring dense multiplexing and real-time performance.
