Open AccessComparison of Time Warping Algorithms for Rail Vehicle Velocity Estimation in Low Speed Scenarios
Author(s) -
Stefan Hensel,
Marin B. Marinov
Publication year - 2017
Publication title -
metrology and measurement systems/metrologia i systemy pomiarowe
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.291
H-Index - 30
eISSN - 2080-9050
pISSN - 0860-8229
DOI - 10.1515/mms-2017-0012
Subject(s) - dynamic time warping , image warping , algorithm , signal (programming language) , computer science , dynamic programming , interval (graph theory) , mathematics , artificial intelligence , combinatorics , programming language
Precise measurement of rail vehicle velocities is an essential prerequisite for the implementation of modern train control systems and the improvement of transportation capacity and logistics. Novel eddy current sensor systems make it possible to estimate velocity by using cross-correlation techniques, which show a decline in precision in areas of high accelerations. This is due to signal distortions within the correlation interval. We propose to overcome these problems by employing algorithms from the field of dynamic programming. In this paper we evaluate the application of correlation optimized warping, an enhanced version of dynamic time warping algorithms, and compare it with the classical algorithm for estimating rail vehicle velocities in areas of high accelerations and decelerations