Self-labelling of tugboat operation using unsupervised machine learning and intensity indicator

Januwar Hadi, Dimitrios Konovessis, Zhi Yung Tay

Research output: Contribution to journalArticlepeer-review

3 Downloads (Pure)


The actual operational data, such as a time sequence of fuel consumption and speed, is usually unlabeled or not associated with a specific activity like tugging or cruising. The operation type is critical for further analysis, as tugging and cruising operations require different fuel and navigation profiles. This paper aims to develop a self-labelling framework for tugboat operation by using unsupervised machine learning and a proposed intensity indicator. The framework considers two sets of data: the positional data and the fuel consumption rate data. The fuel consumption data is obtained from mass flowmeters installed on tugboats, while the positional data are navigational data purchased from marine data aggregators. The developed self-labelling enables ship operators in identifying operations and locations that require heavy fuel consumption andcan be used for further big data analytics and machine learning for fuel consumption prediction when vessel speeds are known.
Original languageEnglish
Article number100082
Number of pages14
JournalMaritime Transport Research
Early online date31 Jan 2023
Publication statusE-pub ahead of print - 31 Jan 2023


  • machine learning
  • self-labelling
  • intensity indicators
  • K-means clustering
  • fuel prediction


Dive into the research topics of 'Self-labelling of tugboat operation using unsupervised machine learning and intensity indicator'. Together they form a unique fingerprint.

Cite this