Advertisement

Exploratory Trajectory Clustering with Distance Geometry

  • Andrew T. WilsonEmail author
  • Mark D. Rintoul
  • Christopher G. Valicka
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 9744)

Abstract

We present here an example of how a large, multi-dimensional unstructured data set, namely aircraft trajectories over the United States, can be analyzed using relatively straightforward unsupervised learning techniques. We begin by adding a rough structure to the trajectory data using the notion of distance geometry. This provides a very generic structure to the data that allows it to be indexed as an n-dimensional vector. We then do a clustering based on the HDBSCAN algorithm to both group flights with similar shapes and find outliers that have a relatively unique shape. Next, we expand the notion of geometric features to more specialized features and demonstrate the power of these features to solve specific problems. Finally, we highlight not just the power of the technique but also the speed and simplicity of the implementation by demonstrating them on very large data sets.

Keywords

Feature Vector Minimum Span Tree Collision Avoidance Anomaly Detection Trajectory Data 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Notes

Acknowledgements

We are grateful to AirNav Systems LLC for providing access to the ASDI data feed. We also applaud Leland McInnes for his highly optimized Python implementation of HDBSCAN.

References

  1. 1.
    Amidan, B., Ferryman, T.: Atypical event and typical pattern detection within complex systems. In: 2005 IEEE Aerospace Conference, pp. 3620–3631, March 2005Google Scholar
  2. 2.
    Annoni, R., Forster, C.: Analysis of aircraft trajectories using fourier descriptors and kernel density estimation. In: 2012 15th International IEEE Conference on Intelligent Transportation Systems (ITSC), pp. 1441–1446, September 2012Google Scholar
  3. 3.
    Bashir, F.I., Khokhar, A.A., Schonfeld, D.: Object trajectory-based activity classification and recognition using Hidden Markov Models. IEEE Trans. Image Process 16, 1912–1919 (2007)MathSciNetCrossRefGoogle Scholar
  4. 4.
    Boem, F., Pellegrino, F.A., Fenu, G., Parisini, T.: Multi-feature trajectory clustering using earth mover’s distance. In: CASE, pp. 310–315. IEEE (2011). http://dblp.uni-trier.de/db/conf/case/case2011.html#BoemPFP11
  5. 5.
    Campello, R.J.G.B., Moulavi, D., Zimek, A., Sander, J.: Hierarchical density estimates for data clustering, visualization, and outlier detection. ACM Trans. Knowl. Discov. Data 10(1), 5:1–5:51 (2015). http://doi.acm.org/10.1145/2733381 CrossRefGoogle Scholar
  6. 6.
    Crippen, G., Havel, T.: Distance geometry and molecular conformation. Chemometrics series. Research Studies Press (1988). http://books.google.com/books?id=5CPwAAAAMAAJ
  7. 7.
    Ester, M., Peter Kriegel, H., Sander, J., Xu, X.: A density-based algorithm for discovering clusters in large spatial databases with noise, pp. 226–231. AAAI Press (1996)Google Scholar
  8. 8.
    Guan, B., Liu, L., Chen, J.: Using relative distance and Hausdorff distance to mine trajectory clusters. TELKOMNIKA Indonesian J. Electr. Eng. 11(1), 115–122 (2013). http://iaesjournal.com/online/index.php/TELKOMNIKA/article/view/1877 Google Scholar
  9. 9.
    Guyon, I., Elisseeff, A.: An introduction to variable and feature selection. J. Mach. Learn. Res. 3, 1157–1182 (2003). http://dl.acm.org/citation.cfm?id=944919.944968 zbMATHGoogle Scholar
  10. 10.
    Kooij, J.F.P., Englebienne, G., Gavrila, D.M.: A non-parametric hierarchical model to discover behavior dynamics from tracks. In: Fitzgibbon, A., Lazebnik, S., Perona, P., Sato, Y., Schmid, C. (eds.) ECCV 2012, Part VI. LNCS, vol. 7577, pp. 270–283. Springer, Heidelberg (2012). http://dx.doi.org/10.1007/978-3-642-33783-3-20 CrossRefGoogle Scholar
  11. 11.
    Lee, J.G., Han, J., Li, X., Gonzalez, H.: Traclass: trajectory classification using hierarchical region-based and trajectory-based clustering. Proc. VLDB Endow. 1(1), 1081–1094 (2008). http://dx.doi.org/10.14778/1453856.1453972 CrossRefGoogle Scholar
  12. 12.
    Nascimento, J., Figueiredo, M., Marques, J.: Trajectory classification using switched dynamical Hidden Markov Models. IEEE Trans. Image Process. 19(5), 1338–1348 (2010)MathSciNetCrossRefGoogle Scholar
  13. 13.
    Pallotta, G., Vespe, M., Bryan, K.: Vessel pattern knowledge discovery from ais data: a framework for anomaly detection and route prediction. Entropy 15(6), 2218 (2013). http://www.mdpi.com/1099-4300/15/6/2218 CrossRefzbMATHGoogle Scholar
  14. 14.
    Patterson, T.A., Thomas, L., Wilcox, C., Ovaskainen, O., Matthiopoulos, J.: State space models of individual animal movement. Trends Ecol. Evol. 23(2), 87–94 (2008). http://www.sciencedirect.com/science/article/pii/S0169534707003588 CrossRefGoogle Scholar
  15. 15.
    Rintoul, M.D., Wilson, A.T.: Trajectory analysis via a geometric feature space approach. Stat. Anal. Data Min. 8(5–6), 287–301 (2015). http://dx.doi.org/10.1002/sam.11287 MathSciNetCrossRefGoogle Scholar
  16. 16.
    Yuan, J., Zheng, Y., Xie, X., Sun, G.: Driving with knowledge from the physical world. In: Proceedings of the 17th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD 2011, pp. 316–324. ACM, New York (2011). http://doi.acm.org/10.1145/2020408.2020462
  17. 17.
    Zheng, Y., Zhang, L., Xie, X., Ma, W.Y.: Mining interesting locations and travel sequences from gps trajectories. In: Proceedings of the 18th International Conference on World Wide Web, WWW 2009, pp. 791–800. ACM, New York (2009). http://doi.acm.org/10.1145/1526709.1526816

Copyright information

© Springer International Publishing Switzerland 2016

Authors and Affiliations

  • Andrew T. Wilson
    • 1
    Email author
  • Mark D. Rintoul
    • 1
  • Christopher G. Valicka
    • 1
  1. 1.Sandia National LaboratoriesAlbuquerqueUSA

Personalised recommendations