A Subspace Method for Time Series Anomaly Detection in Cyber-Physical Systems

TL;DR

This paper introduces a fast, low-cost subspace method for detecting anomalies in time series data from cyber-physical systems, supported by theoretical foundations and practical algorithms.

Contribution

The paper proposes a novel subspace-based approach for real-time anomaly detection in sensor signals, with theoretical analysis and a prototype algorithm.

Findings

Method achieves low computational cost

Effective in detecting anomalies in dynamical systems

Provides computational tools for practical implementation

Abstract

Time series anomaly detection is an important process for system monitoring and model switching, among other applications in cyber-physical systems. In this document, we present a fast subspace method for time series anomaly detection, with a relatively low computational cost, that has been designed for anomaly detection in real sensor signals corresponding to dynamical systems. We also present some general results corresponding to the theoretical foundations of our method, together with a prototypical algorithm to for time series anomaly detection. Some numerical examples corresponding to applications of the prototypical algorithm are presented, and some computational tools based on the theory and algorithms presented in this paper, are provided.