| ICS and the Florida Institute of Technology
Announce NASA STTR Award
Melbourne, FL., October 11, 2001 -- Interface & Control Systems (ICS) announced today that it is working with Florida Tech. to add support for Machine Learning technology to its flagship SCL product suite, funded by a NASA Small Business Technology Transfer (STTR) award. The effort is entitled “Automated Learning for Real-Time Expert System in Monitoring and Control”. Until
now, the Aerospace Industry has had no example of cooperation between a
Rule-Based Expert System and an Adaptive Learning System in a production,
launch or flight environment. Both
methodologies have very strong capabilities and some limitations.
For example, SCL’s sequence-based executive makes it quite simple for
an engineer to specify the steps to be taken while in nominal operational
states, and its rule-based expert system makes it easy to describe procedures
for recovering to a nominal state when off-nominal events occur.
But each of those states and rules must be defined to SCL.
Traditionally, expert system rules and data definitions have been
hand-crafted by engineers, a tedious and costly process.
ICS
and Dr. Philip Chan of Florida Tech. believe that by employing adaptive learning technologies, expert
system rule sets and data definitions can be initially derived and
continuously refined from archived test data and/or real-time data streams,
thus providing a continuous improvement in monitoring capabilities.
Data mining algorithms will be used to analyze recorded data and
extract relational information using clustering techniques for locating ranges
of measurements in normal operations and processing high-dimensional data more
efficiently. This relational
information will be used to facilitate SCL Expert System rule construction and
incremental updates. This
technology will allow a quantum leap in accuracy and usability for advisory
and fault detection systems. Florida Tech.’s
research into adaptive learning and data mining systems holds great promise
for non-invasive applications to mine knowledge from data streams.
This
technology will be especially effective when retro-applied to legacy systems
or to any systems that produce a well-defined data stream which can be used
for data mining. ICS is a product development and engineering services company specializing in the development of real-time, embedded, and autonomous command and control software systems. Services include software systems engineering, integration and test, and project and software life cycle management. ICS is a privately held company with offices in Columbia, MD, Melbourne, FL, and Boulder, CO. See more information on the Automated Machine Learning at: http://www.interfacecontrol.com/news/AML_STTR.htm Contact information:
 |
