Abstract

This article introduces an interrogation algorithm designed for low-cost tunable laser-based fiber Bragg grating (FBG) interrogation systems, specifically targeting general-purpose audio sensing. The algorithm leverages the inherent continuity and regularity present in acoustic signals by employing a prediction filter to forecast future samples. The algorithm determines the scan region based on the prediction and then acquires the narrowest optical bands that encompass solely the necessary information. Peak detection is performed on a subwindow of adaptive size. Compared with conventional full-band acquisition, the proposed algorithm has a much shorter laser engine scan path thus the overall sampling rate is augmented. Numerical simulations demonstrate a signal-to-noise ratio of 76.9 dB and an interrogation error of $\pm$ 9.747 pm (99.7% confidence) using properties of real-world optoelectronic devices and diversified audio samples. The algorithm presents consistent performance across audio content and a significant sampling rate improvement of maximum 40 times compared to conventional full-band scanning. The algorithm offers flexible configuration, hardware efficiency, and scalability for dynamic FBG sensing applications.