The locomotive diagnostic system is comprised of three subsystems: a diagnostic recorder unit, a train-borne recorder, and a graphical user interface (GUI) for data display and evaluation. The diagnostic unit collects data from sensors located in the locomotive, scales them and displays them in the GUI. Machine learning algorithms are then applied to the processed data, leveraging historical data and expert knowledge to identify patterns and correlations between different parameters.

Additionally, the algorithms estimate the remaining useful life of components, enabling predictive maintenance. The train-borne recorder captures audio recordings of the communication between the locomotive driver and the railway operator. This data is also displayed in the GUI, sorted according to time and GPS location, making it easy to determine who is responsible in the event of an accident.