Fast and Accurate Measurement in BOTDA Fiber Sensor Through the Application of Filtering Techniques in Frequency and Time Domains

This study introduces a new approach to enhance the measurement speed in the suggested system-on-chip (SoC) by incorporating quadratic least mean square (LMS) filtering with moving average (MA) and moving median (MM) filters at a low-averaging stage. The use of MA and MM filters in the frequency and time domains yields a significant enhancement in the reliability and accuracy of the Brillouin frequency shift (BFS). Our method, as demonstrated in this study, exhibits a significantly reduced latency, which ranges from 14 to 1563 times faster than Brillouin optical time-domain analysis (BOTDA) systems that rely on an oscilloscope and a computer for data processing, and the BOTDA system utilizing SoC with LMS filtering alone. Compared with LMS curve fitting alone, the implementation of this technique greatly enhances accuracy with low averages by 84%. These filters enable the identification of minor disruptions along the cable. The proposed method successfully detects small variations in Brillouin frequency while maintaining the spatial resolution of sensors. The deployed filters result in a spatial resolution ranging from 1.6 to 1.5 m, which demonstrates greater constancy than the normal range of 2–1.6 m. This study successfully employs filtering techniques in the frequency and time domains by incorporating low averaging and a broad dynamic range. The use of SoC technology in BOTDA improves measurement speed, accuracy, and the preservation of spatial resolution constancy. However, the measurable fiber length is limited due to ac coupling on the analog-to-digital converter board. The algorithm and dataset are publicly available at GitHub.