Abstract
Existing data mining approaches to complex systems anomaly detection use uni-variate and/or multi-variate statistical hypothesis testing to assign anomaly scores to data streams associated with system components. The former approach assumes statistical independence of individual components, while the latter assumes substantial global systemic correlation. As a compromise between these two epistemological extremes, we present a data-mining approach hybridizing existing statistical techniques with theorem-proving methods to create a novel algorithm for anomaly detection, diagnosis and control in complex systems. Our algorithm takes sensor inputs from physical sensors providing system subcomponent performance data and outputs (i) a global systemic risk indicator and (ii) possible diagnosis hypotheses. We present results on three different systems, and in comparison with current state-of-the-art fault detection algorithms to demonstrate the viability of our approach. We find that our algorithm proves robust towards increased data dimensionality in contrast with existing clustering-based fault detection methods and can also detect contextual faults that are undetectable using existing statistical techniques.
Original language | English |
---|---|
Title of host publication | Annual Conference of the Prognostics and Health Management Society, PHM 2010 |
Publisher | Prognostics and Health Management Society |
ISBN (Print) | 9781936263011 |
Publication status | Published - 2010 |
Externally published | Yes |
Event | Annual Conference of the Prognostics and Health Management Society, PHM 2010 - Portland Duration: 13 Oct 2010 → 16 Oct 2010 |
Other
Other | Annual Conference of the Prognostics and Health Management Society, PHM 2010 |
---|---|
City | Portland |
Period | 13/10/10 → 16/10/10 |
Fingerprint
ASJC Scopus subject areas
- Electrical and Electronic Engineering
- Health Information Management
- Computer Science Applications
- Software
Cite this
A hybrid-logic approach towards fault detection in complex cyber-physical systems. / Srivastava, Nisheeth; Srivastava, Jaideep.
Annual Conference of the Prognostics and Health Management Society, PHM 2010. Prognostics and Health Management Society, 2010.Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - A hybrid-logic approach towards fault detection in complex cyber-physical systems
AU - Srivastava, Nisheeth
AU - Srivastava, Jaideep
PY - 2010
Y1 - 2010
N2 - Existing data mining approaches to complex systems anomaly detection use uni-variate and/or multi-variate statistical hypothesis testing to assign anomaly scores to data streams associated with system components. The former approach assumes statistical independence of individual components, while the latter assumes substantial global systemic correlation. As a compromise between these two epistemological extremes, we present a data-mining approach hybridizing existing statistical techniques with theorem-proving methods to create a novel algorithm for anomaly detection, diagnosis and control in complex systems. Our algorithm takes sensor inputs from physical sensors providing system subcomponent performance data and outputs (i) a global systemic risk indicator and (ii) possible diagnosis hypotheses. We present results on three different systems, and in comparison with current state-of-the-art fault detection algorithms to demonstrate the viability of our approach. We find that our algorithm proves robust towards increased data dimensionality in contrast with existing clustering-based fault detection methods and can also detect contextual faults that are undetectable using existing statistical techniques.
AB - Existing data mining approaches to complex systems anomaly detection use uni-variate and/or multi-variate statistical hypothesis testing to assign anomaly scores to data streams associated with system components. The former approach assumes statistical independence of individual components, while the latter assumes substantial global systemic correlation. As a compromise between these two epistemological extremes, we present a data-mining approach hybridizing existing statistical techniques with theorem-proving methods to create a novel algorithm for anomaly detection, diagnosis and control in complex systems. Our algorithm takes sensor inputs from physical sensors providing system subcomponent performance data and outputs (i) a global systemic risk indicator and (ii) possible diagnosis hypotheses. We present results on three different systems, and in comparison with current state-of-the-art fault detection algorithms to demonstrate the viability of our approach. We find that our algorithm proves robust towards increased data dimensionality in contrast with existing clustering-based fault detection methods and can also detect contextual faults that are undetectable using existing statistical techniques.
UR - http://www.scopus.com/inward/record.url?scp=84920518109&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84920518109&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:84920518109
SN - 9781936263011
BT - Annual Conference of the Prognostics and Health Management Society, PHM 2010
PB - Prognostics and Health Management Society
ER -