Vibration analysis and neural networks for proactive railway track monitoring using fiber optic sensors

Abstract

The safety and efficiency of rail transport directly depend on the integrity of the track infrastructure. However, traditional monitoring methods, such as visual inspections, non-destructive testing, and rail loops, are mostly reactive or periodic. They leave significant time gaps during which critical defects can arise and develop, and their effectiveness is limited by subjectivity, high cost, and vulnerability to external factors. This creates an urgent need for innovative automated systems capable of providing continuous, highly accurate, and, most importantly, proactive monitoring of the condition of railway tracks in real time.