Conversion of time-sampled data of track irregularity to distance-sampled data by dynamic programming based elastic matching

摘要

In this paper we give a method of conversion of a time-sampled data to a distance-sampled data for a same one data of railway track irregularity. On the railway this conversion is sometimes required. A waveform of a time-sampled sequence will look like stretched and shortened, if viewed on the distance-axis, relatively comparing with that of a distance-sampled sequence. A correspondence between two waveforms takes a form of warped mapping relation. An extent of such warping behavior depends on a variation in a running velocity of the track recording car. The waveforms of these two kinds of digital sequences originally should correspond with each other. By applying a method of dynamic programming based elastic matching to this problem an efficient conversion of a time-sampled data to a distance-sampled data is made possible without using any information of the running velocity of the car.