Time series anomaly detection for heavy-duty vehicles

Title Time series anomaly detection for heavy-duty vehicles
Summary Detecting anomalies in multivariate time series data collected from vehicle operations
TimeFrame Fall 2023
References Ruiz, C., Menasalvas, E., & Spiliopoulou, M. (2009). C-denstream: Using domain knowledge on a data stream. In Discovery Science: 12th International Conference, DS 2009, Porto, Portugal, October 3-5, 2009 12 (pp. 287-301). Springer Berlin Heidelberg.

Cao, F., Estert, M., Qian, W., & Zhou, A. (2006, April). Density-based clustering over an evolving data stream with noise. In Proceedings of the 2006 SIAM international conference on data mining (pp. 328-339). Society for industrial and applied mathematics.

Fan, Y., Nowaczyk, S., & Antonelo, E. A. (2016). Predicting air compressor failures with echo state networks. In PHM Society European Conference (Vol. 3, No. 1).

Fan, Y., Nowaczyk, S., & Rögnvaldsson, T. (2020). Transfer learning for remaining useful life prediction based on consensus self-organizing models. Reliability Engineering & System Safety, 203, 107098.

Ahmad, S., Lavin, A., Purdy, S., & Agha, Z. (2017). Unsupervised real-time anomaly detection for streaming data. Neurocomputing, 262, 134-147.

Lavin, A., & Ahmad, S. (2015, December). Evaluating real-time anomaly detection algorithms--the Numenta anomaly benchmark. In 2015 IEEE 14th international conference on machine learning and applications (ICMLA) (pp. 38-44). IEEE.

Wu, K., Zhang, K., Fan, W., Edwards, A., & Philip, S. Y. (2014, December). Rs-forest: A rapid density estimator for streaming anomaly detection. In 2014 IEEE international conference on data mining (pp. 600-609). IEEE.

Hendrickx, K., Meert, W., Mollet, Y., Gyselinck, J., Cornelis, B., Gryllias, K., & Davis, J. (2020). A general anomaly detection framework for fleet-based condition monitoring of machines. Mechanical Systems and Signal Processing, 139, 106585.

Supervisor Yuantao Fan, Hamid Sarmadi
Status Open

This project will explore and develop algorithms to detect anomalies in multi-variate time series data, collected from heavy-duty vehicles. The idea is to identify different operating modes using stream clustering algorithms and discover anomalies in each cluster, i.e. combing DenStream and Consensus self-organizing models methods (COSMO) for anomaly detection. The goal is to discover anomalous events in advance to enhance vehicle operation safety. The project includes: i) proposing a streaming anomaly detection framework; ii) finding suitable data representations that can capture key characteristics in vehicle operation; and iii) evaluating the proposed approach on a real-world dataset. This work is a collaboration with industry.