Two algorithms for the sorting of unknown train vibration signals into freight and passenger train categories

Calum Sharp , David Waddington , James Woodcock , Gennaro Sica , Eulalia Peris , Andy Moorhouse

The human response to railway noise is well researched; however, there is a need to further research the human response to freight railway vibration. To facilitate this, two algorithms have been constructed with the aim of sorting unknown train vibration signals into freight and passenger train categories so that they can be further analysed. 307 known train vibration signals measured close to the railway were analysed to determine which signal properties, if any, are identifiably different for freight and passenger train vibration signals. These data were collected within the Defra funded UK study “Human Response to Vibration in Residential Environments” conducted by the University of Salford. Several signal properties were found to be statistically significantly different for freight and passenger train vibration signals, all of which relate either to the duration of the signal event or its frequency content. These differences were used to successfully create two algorithms that are capable of sorting unknown train signals into freight and passenger train categories at a relatively high level of accuracy. The methodology used in creating the algorithms, their level of accuracy and recommendations for their use will be presented in this paper.

Mots clés: Railway; Freight; Noise; Vibration

Domaines: Acoustique [physics.class-ph]

Calum Sharp, David Waddington, James Woodcock, Gennaro Sica, Eulalia Peris, et al.. Two algorithms for the sorting of unknown train vibration signals into freight and passenger train categories. Acoustics 2012, Apr 2012, Nantes, France. ⟨hal-00810694⟩