The article discusses the implementation potential of the Artificial Intelligence methods in monitoring and analysis of permanent way condition monitoring.
Regular inspection of railway track conditions is crucial for maintaining safe and reliable train operations. The diagnostic track recording vehicles and trolleys collect voluminous accurate information on track and turnouts' safety and functional parameters. The traditional analysis of this data made by human experts only turns out to be less efficient and prone to human error than automated analyses.
Research into artificial intelligence yielded methods to carry out tasks previously considered too complex to be done without human intervention. Most inspection data can be analysed automatically, be it track and turnout geometry readings and video inspection information. If required, unique annotation overlays and reporting procedures can be applied to provide instantaneous results.
Information collected by the test vehicles provides diagnostic data, which the diagnostic software can analyse on the intelligent platform. This intelligent platform can use various Artificial Intelligence tools like expert systems, intelligent agents continuously browsing the diagnostic results database, Genetic Algorithms, Neural Networks, or Bayesian framework as a self-learning system.
The automated and unbiased analysis results make sound maintenance decisions possible. Such an approach makes allocating the limited budgets and resources possible with various priorities to optimise the amount of investment required to keep the assets in good health. Efficient maintenance planning has become possible with maintenance work schedules, work order generation, work maintenance support and others, categorising the track and turnout quality based on the collected information.

Artificial Intelligence; Intelligent Agents; track and turnout condition diagnostics