Partnering with leading researchers in the field of data mining (Dr. Philip Chan and Dr. Matt Mahoney of Florida Tech), ICS has developed a time-series data mining tool called sensorMiner. The goal was to build a sophisticated data mining toolset that uses past performance data to build a human-readable model to ultimately provide real-time fault detection capability. SensorMiner learns from past performance data and automatically generates a monitoring system. Conventional technology has made this process both a labor-intensive and error-prone task often misunderstood, even by experts. SensorMiner will correlate seemingly unrelated data automatically and provide rich graphical feedback to the user.

The sensorMiner toolset is a unique approach to data mining using a combination of time warp, cluster, rule induction, Euclidean error, and state machine technologies resulting in a sophisticated temporal machine learner for real-time anomaly detection. Using time-series anomaly detection, the toolset trains a model on known good data, estimates the probability distribution, and assigns a likelihood-based score to new sensor data.

“The ability to automatically make predictions or help people make decisions faster and more accurately - in real-time - has far-reaching implications that spread across industry boundaries," said Brian Buckley, VP of Marketing for ICS. "We designed sensorMiner to identify anomalies in a wide variety of systems".

Spearheading the sensorMiner development effort is Walter Schiefele, principle scientist for ICS. “Many artificial intelligence programs yield models that are not understandable by humans. Our approach was to generate models that are easily readable and graphically displayed. People can visualize the rules. More importantly, domain experts can manually trim or adjust rules in the final product.”

The sensorMiner toolset is a commercially available product that (1) learns and generalizes current traces; (2) formulates a concise model of the known good data signatures; (3) applies the generated model to on-line, real-time data; (4) detects on-line (real-time) anomalies and/or deviations from the model; and (5) reports anomalies to the operator with explicit information as to where the signal deviated and exactly what constraints were violated. ICS hopes to penetrate existing Control System and Test Set markets that require more sophisticated real-time anomaly detection and analysis.

Potential customers wanting to evaluate the sensorMiner toolset must be qualified by contacting ICS at 321-723-0399, ask for Brian Buckley. Alternatively, contact us by e-mailing info@interfacecontrol.com