Fiber-Bragg-grating-based sensing for dynamic strain and speech recognition

In this work, we report on the use of fiber Bragg grating (FBG) with post-processing techniques to enhance speech recognition quality. Our approach integrates FBGs with signal processing methods, including DC removal, bandpass filtering, and least mean square (LMS) filtering, to improve the quality of recorded speech signals. We then used open speech recognition services and the open Harvard dataset to achieve a word error rate (WER) of 26% and a Levenshtein distance (LD) of 50. These results highlight the potential of FBG-based systems for applications requiring robust speech recognition, especially in environments where traditional microphones are ineffective.