Vibration Signals Modeling of Train Based on Distributed Optical Fiber Sensing

This paper develops innovative train vibration analysis methods using Distributed Acoustic Sensing (DAS) for intelligent rail systems. While conventional systems face limited spatial coverage and electromagnetic interference, DAS enables full-range monitoring with immunity to interference and simplified installation, proving ideal for long-distance vibration detection. However, its high sensitivity introduces environmental noise, complicating signal extraction. To address this, we propose a framework that first separates signal-noise subspaces through matrix decomposition and source estimation, then enhances features via iterative singular value thresholding, and finally reconstructs principal frequency components using a multi- component sinusoidal model. This approach advances vibration processing while preserving DAS advantages, offering insights for intelligent rail development.