FDM φ-OTDR with 1.7-m spatial resolution using an unsupervised aliasing interference suppressing algorithm

Frequency division multiplexing (FDM) φ-OTDR naturally features a wide frequency bandwidth, excellent interference fading suppression, and high system robustness. These features are rarely available together on other φ-OTDR systems, including those using chirped-pulse/frequency-sweeping schemes. However, achieving finer spatial resolution necessitates the use of shorter-duration pulses, which introduces aliasing interference in the frequency-divided traces and typically degrades spatial resolution to beyond 5 m. To resolve this issue, we propose an unsupervised dictionary learning (u-DIC) algorithm that facilitates the use of ultrashort pulses for fine spatial resolution while effectively suppressing aliasing interference. The dictionary is adaptively learned from noisy data and requires no prior noise model or labeled dataset, ensuring robustness in diverse operating environments. Experimentally, this algorithm achieves distributed acoustic signal recovery with a spatial resolution of 1.7 m over an 8-km sensing range. The accurate recovery of a square-wave perturbation confirms the complete preservation of signal characteristics. This algorithm will help FDM φ-OTDR find new foreground in high-resolution distributed acoustic sensing, like robotics and aeronautic structures.