Welcome to the National Transport Library Catalogue

Normal view MARC view

Design and implementation of a maritime traffic modeling and anomaly detection method Osekowska, Ewa

By: Series: Blekinge Institute of Technology licentiate dissertation series ; 2014:09Publication details: Karlskrona Blekinge Institute of Technology, 2014Description: 129 sISBN:
  • 9789172952911
Subject(s): Online resources: Dissertation note: Licentiatavhandling Karlskrona : Blekinge Institute of Technology, 2014 Abstract: Nowadays ships are usually equipped with an array of marine instruments, one of which is an Automatic Identification System (AIS) transponder. The availability of the global AIS ship tracking data opened the possibility to develop maritime security far beyond simple collision prevention. The research work summarized in this thesis explores this opportunity, with the aim of developing an intuitive and comprehensible method for traffic modeling and anomaly detection in the maritime domain. The novelty of the method lies in employing the technique of artificial potential fields for traffic pattern extraction. The general idea is for the potentials to represent typical patterns of vessels’ behaviors. A conflict between potentials, which have been observed in the past, and the potential of a vessel currently in motion, indicates an anomaly. The proposed potential field based method has been examined using a web-based anomaly detection system STRAND (Seafaring Transport ANomaly Detection) implemented for this study. Its applicability has been demonstrated in several publications, examining its scalability, modeling capabilities and detection performance. The experimental investigations led to identifying optimal detection resolution for different traffic areas (open sea, harbor and river), and extracting traffic rules, e.g., with regard to speed limits and course rules. Furthermore, the map-based display of modeled traffic patterns and detection cases has been analyzed, using several demonstrative cases. The massive AIS dataset collected for this study provides an abundance of challenges for future investigations.
Item type: Licentiate thesis
No physical items for this record

Licentiatavhandling Karlskrona : Blekinge Institute of Technology, 2014

Nowadays ships are usually equipped with an array of marine instruments, one of which is an Automatic Identification System (AIS) transponder. The availability of the global AIS ship tracking data opened the possibility to develop maritime security far beyond simple collision prevention. The research work summarized in this thesis explores this opportunity, with the aim of developing an intuitive and comprehensible method for traffic modeling and anomaly detection in the maritime domain. The novelty of the method lies in employing the technique of artificial potential fields for traffic pattern extraction. The general idea is for the potentials to represent typical patterns of vessels’ behaviors. A conflict between potentials, which have been observed in the past, and the potential of a vessel currently in motion, indicates an anomaly. The proposed potential field based method has been examined using a web-based anomaly detection system STRAND (Seafaring Transport ANomaly Detection) implemented for this study. Its applicability has been demonstrated in several publications, examining its scalability, modeling capabilities and detection performance. The experimental investigations led to identifying optimal detection resolution for different traffic areas (open sea, harbor and river), and extracting traffic rules, e.g., with regard to speed limits and course rules. Furthermore, the map-based display of modeled traffic patterns and detection cases has been analyzed, using several demonstrative cases. The massive AIS dataset collected for this study provides an abundance of challenges for future investigations.