Casola et al. [31], Flammini et al. [48] and Maly et al.
[82] proposed to integrate data from a range of heterogeneous
sensors for more holistic condition monitoring. Maly et al.
[82] performed hot box detection, flat wheel detection, dynamic
scaling, and loading gauge detection using commercially available
sensors, whereas Casola et al. [31] analyzed temperature,
humidity, and vibrations among others (the specific sensor
types are not stated). The sensors are grouped into units, which
operate independently. If a unit detects a faulty train, then it
alerts the control center and provides detailed information about
the detected irregularity