The Recent Applications of Machine Learning in Rail Track Maintenance: A Survey

Muhammad Chenariyan Nakhaee, Djoerd Hiemstra, Mariëlle Stoelinga, Martijn van Noort

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

42 Citations (Scopus)
805 Downloads (Pure)


Railway systems play a vital role in the world’s economy and movement of goods and people. Rail tracks are one of the most critical components needed for the uninterrupted operation of railway systems. However, environmental conditions or mechanical forces can accelerate the degradation process of rail tracks. Any fault in rail tracks can incur enormous costs or even results in disastrous incidents such as train derailment. Over the past few years, the research community has adopted the use of machine learning (ML) algorithms for diagnosis and prognosis of rail defects in order to help the railway industry to carry out timely responses to failures. In this paper, we review the existing literature on the state-of-the-art machine learning-based approaches used in different rail track maintenance tasks. As one of our main contributions, we also provide a taxonomy to classify the existing literature based on types of methods and types of data. Moreover, we present the shortcomings of current techniques and discuss what research community and rail industry can do to address these issues. Finally, we conclude with a list of recommended directions for future research in the field.

Original languageEnglish
Title of host publicationReliability, Safety, and Security of Railway Systems. Modelling, Analysis, Verification, and Certification
Subtitle of host publicationThird International Conference, RSSRail 2019, Lille, France, June 4-6, 2019, Proceedings
EditorsSimon Collart-Dutilleul, Thierry Lecomte, Alexander B. Romanovsky
Place of PublicationCham
Number of pages15
ISBN (Electronic)978-3-030-18744-6
ISBN (Print)978-3-030-18743-9
Publication statusPublished - 24 Apr 2019

Publication series

NameLecture Notes in Computer Science
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


  • Deep learning
  • Machine learning
  • Maintenance
  • Rail track
  • 2023 OA procedure


Dive into the research topics of 'The Recent Applications of Machine Learning in Rail Track Maintenance: A Survey'. Together they form a unique fingerprint.

Cite this