TITLE: Artificial intelligence in predictive maintenance and fault detection on railway tracks
AUTHORS: Amila Keško, Hamdo Puška
ABSTRACT: Due to the high demand for railway transportation, which is one of the primary modes of passenger and freight transport, we encounter challenges in forming maintenance plans for vehicles and tracks and coordinating train movements on the railway network. Numerous maintenance methods and techniques have been developed, one of which is predictive maintenance. In fault detection through predictive maintenance, the application of artificial intelligence plays a significant role. Its capabilities are particularly exploited in diagnosing faults on railway tracks. Previous research has shown that artificial intelligence is used in predicting train delays, scheduling timetables, enhancing traffic safety, as well as in maintenance and diagnosis of potential faults that may compromise safety. Techniques such as machine learning, deep learning, neural networks, and the Internet of Things are employed for data collection and processing through smart sensors. By collecting and analyzing data, AI has developed innovative models capable of predicting potential faults and detecting current damages on the track. This paper will investigate advanced techniques based on deep neural networks (DNNs) and convolutional neural networks (CNNs) to support real-time fault identification via smart sensors and cameras. The research results contribute to better railway traffic organization, simplified maintenance planning, and minimizing unplanned train traffic disruptions.
KEYWORDS: artificial intelligence, railway track, predictive maintenance, machine learning
PAGES: 127-134
DOI: