An application of kalman filtering estimation to the condition monitoring of the geometric quality of the railway track

In this paper, an application of the treatment of dynamic datasets of a railway vehicle, as previously described by the same authors in Reference [1], is presented. In the preceding work, the proposed track condition monitoring system was designed to be implemented on board of train being operated in standard revenue service. Through the use of Kalman filter state estimator it was possible to replace some direct measures with their estimation. This paper provides a more comprehensive framework for the study, by investigating the observability of the system. Moreover, the proposed methodology, previously validated exploiting a multi-body simulator, is herein applied to real data collected during measurements campaign on Italian railway network.