Harmonic peak locations based on the twice fast Fourier transform algorithm in a Sagnac intrusion sensing system

Null frequencies in the spectrum of a phase difference signal are used for intrusion localization since their values are associated with the position of intrusion in a Sagnac interferometer. However, searching the null frequencies individually is troublesome and unreliable for localization. In this paper, an algorithm which directly captures the interval of null frequencies is proposed to achieve a simpler and more robust intrusion localization. By applying the modified spectrum to a second fast Fourier transform, the interval responds as multiple peaks in the resulting location curve. After determining the values of the base peak and its harmonic peaks, the intrusion's distance can be localized. Intrusions are applied at a position of 130 km in a sensing fiber. Due to the resolution increase as peak order, localization performance of 2nd harmonic peak is 25 m, which is 34% better than the base peaks. By calculating the average of the localization results by all remarkable peaks, localization performance improves by 47% in laboratory experiments and by 24% in practical applications over the previous single-peak localization algorithm.