In the case of a railway wagon running on track, irregularities in the track formation can lead to complex dynamic interactions. Experience has shown that track geometry alone is not a good predictor of vehicle response and that adverse vehicle responses can occur on track sections with geometry measurements that do not trigger maintenance actions. This paper describes a “health card” which is an autonomous solar powered device that can be placed on rolling stock to capture vehicle responses. Health cards make use of an algorithm which autonomously analyses signals from accelerometers mounted on the wagon body, to detect and classify the dynamic interaction of the track and the rail vehicle. The algorithm presented resolves the wagon body motions into five degrees of freedom, applies Spectrogram techniques to obtain a time-frequency representation, and autonomously analyses the representation to detect and classify transient dynamic events from wagon body motions. Although sensing at the wagon body provides a limited measurement of the wagon system, the results of this research is useful for systems requiring minimal or low cost installation. The algorithm is developed and validated using field data including accelerations measured at strategic points on the wagon body and the bogies.