Pattern of life model parameterization for exploitation in Command and Control systems – Methodology report part I: Target motion model and formalization

PDF

Authors
  1. Millefiori, L.M.
  2. Braca, P.
  3. Horn, S.
Corporate Authors
Defence Research and Development Canada, Centre for Operational Research and Analysis, Ottawa ON (CAN)
Abstract
Extracting valuable information from large spatio-temporal datasets requires innovative approaches that can efficiently deal with large amounts of data and, at the same time, effectively reveal the underlying structure of the data, in order to provide useful information to the decision making process. Innovative knowledge discovery techniques have been developed which use a stochas-tic meanreverting modeling of the ships motion to reveal the underlying graphical structure of maritime traffic. The generated knowledge enables numerous possibil-ities, from graph-based multi-edge prediction to anomaly detection techniques, to ship routing optimization. Altogether, the topics covered in this report represent the theoretical framework that is required for the development of knowledge discovery techniques able to reveal the underlying graph structure of maritime traffic, which are documented in the companion report—Part II. This report—Part I—documents the formalization of the ship motion model, moti-vating its use over other conventional models. Procedures to estimate the process parameters are provided and its use for long-term prediction and data association is investigated. The main limitation of this model, its applicability to non-maneuvering targets only, is also overcome by formalizing an augmented version of the model that fits the case of a vessel navigating by waypoints. Realworld data sets are used to show the potential of the developed techniques in cases of pract

Il y a un résumé en français ici.

Keywords
pattern of life;Machine Learning;Situational Awareness
Report Number
DRDC-RDDC-2019-R058 — Scientific Report
Date of publication
01 Sep 2019
Number of Pages
41
DSTKIM No
CA049936
CANDIS No
810785
Format(s):
Electronic Document(PDF)

Permanent link

Document 1 of 1

Date modified: