Spatio-temporal Anomaly Detection in Intelligent Transportation Systems

With the increased availability of real-time data streams in different domains comes the opportunity of using these data to provide valuable insights into the performance of the systems generating such data. In this paper, the authors are proposing an anomaly detection method to be applied on road traffic data in intelligent transportation systems. The proposed scheme is based on multi-channel singular spectrum analysis (MSSA), and aims to characterize the spatio-temporal properties of the transportation network. By simultaneously analyzing the spatial and temporal attributes of the network, the proposed anomaly detection scheme is able to detect contextual and collective anomalies that are otherwise undetectable using only spatial or temporal anomaly detection techniques. This is indicated through the results shown in the experiments and results section.

Language

  • English

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 01712333
  • Record Type: Publication
  • Files: TRIS
  • Created Date: May 24 2019 5:31PM