Self-calibrating models for dynamic monitoring and diagnosis
- Kuipers, Benjamin
- Jan 1, 1994.
- Physical Description:
- 1 electronic document
- Restrictions on Access:
- Unclassified, Unlimited, Publicly available.
Free-to-read Unrestricted online access
- The present goal in qualitative reasoning is to develop methods for automatically building qualitative and semiquantitative models of dynamic systems and to use them for monitoring and fault diagnosis. The qualitative approach to modeling provides a guarantee of coverage while our semiquantitative methods support convergence toward a numerical model as observations are accumulated. We have developed and applied methods for automatic creation of qualitative models, developed two methods for obtaining tractable results on problems that were previously intractable for qualitative simulation, and developed more powerful methods for learning semiquantitative models from observations and deriving semiquantitative predictions from them. With these advances, qualitative reasoning comes significantly closer to realizing its aims as a practical engineering method.
- NASA Technical Reports Server (NTRS) Collection.
- Document ID: 19940029546.
Accession ID: 94N34052.
NASA. Johnson Space Center, The Seventh Annual Workshop on Space Operations Applications and Research (SOAR 1993), Volume 1; p 245-252.
- No Copyright.
View MARC record | catkey: 15658343