The Outlier Detection Via Estimating Clusters (ODVEC) software provides an efficient method for real-time or offline analysis of multivariate sensor data for use in anomaly detection, fault detection, and system health monitoring. ODVEC uses models automatically derived from archived system data to identify unusual, out-of-family data samples (outliers) that indicate possible system failure or degradation. It employs novel techniques to efficiently calculate the degree of deviation of current system behavior from a range of previous, similar nominal operations.

Automatic system health monitoring can significantly benefit from comparison of system sensor data to an accurate characterization or model of expected system behavior. The Inductive Monitoring System (IMS) was motivated by the difficulty of producing detailed health monitoring and diagnostic models of system components due to complexity or unavailability of design information. One part of IMS, the learning algorithm, automatically extracts system models from archived system data. The IMS monitoring algorithm compares new system data to that model to find unusual data points. Like IMS monitor, ODVEC uses system models produced by IMS learn or similar algorithms. But ODVEC significantly augments the capability of IMS monitor by comparing system data vectors to multiple data samples in the model, rather than just the single best matching data point.

By examining multiple nearby data points, ODVEC also has the ability to identify outlier points within a single data set, a feature not currently available in the IMS monitor algorithm. Additionally, ODVEC provides a method to analyze the model to determine the statistical distribution of the data points used to construct the model. This allows ODVEC monitoring results to be expressed relative to common statistical measures, such as standard deviation. This feature makes the ODVEC output more consistent across systems and easier to interpret than IMS monitor results, which are not standardized and are influenced by system-specific features of the model. ODVEC uses a cluster storage, indexing, and retrieval scheme similar to the IMS monitor techniques.

This work was done by David L. Iverson of Ames Research Center. NASA invites companies to inquire about partnering opportunities and licensing this patented technology. Contact the Ames Technology Partnerships Office at 1-855-627-2249 or This email address is being protected from spambots. You need JavaScript enabled to view it.. Refer to ARC-16467-1.

NASA Tech Briefs Magazine

This article first appeared in the February, 2016 issue of NASA Tech Briefs Magazine.

Read more articles from this issue here.

Read more articles from the archives here.