Detection of anomalies in stationary signals based on fluctuation-dissipation principle

Issue title: Proceedings from the 17th International Symposium on Applied Electromagnetics and Mechanics (ISEM 2015)

Guest editors: Fumio Kojima, Futoshi Kobayashi and Hiroyuki Nakamoto

Article type: Research Article

Authors: Honda, Satoshi^{a; *} | Yonekura, Hiromitsu^b

Affiliations: [a] Department of Applied Physics & Physico-Informatics, Keio University, Yokohama, Japan | [b] MHPS Control Systems Co., Ltd., Yokohama, Japan

Correspondence: [*] Corresponding author: Satoshi Honda, Department of Applied Physics & Physico-Informatics, Keio University, Hiyoshi 3-14-1, Kohoku, Yokohama 223-8522, Japan. E-mail:honda@appi.keio.ac.jp

Abstract: This paper describes the detection method of the anomalies in signals proposed by Okabe et al. based on their KM_2O-Langevin framework; breakdown points of the stationarity can be detected to give the sign of anomalies. The method was applied to the sound signals to monitor the state of a generator engine and showed its feasibility.

Keywords: Anomaly detection, stationarity test, fluctuation-dissipation principle

DOI: 10.3233/JAE-162102

Journal: International Journal of Applied Electromagnetics and Mechanics, vol. 52, no. 1-2, pp. 55-60, 2016