Spacecraft Environmental Anomalies (SEA-ES) is an expert system developed by Aerospace Corporation, for use in the diagnosis of satellite anomalies caused by the space environment. Clustering this rule base with the MVP-CA tool helped expose some inconsistencies as well as redundancies across rules in this knowledge base. Clustering also identified rule groups with functionalities that could be used across missions. These groups consisted of rules dealing with defining the orbits of the satellite, the local plasma and radiation environment, satellite-exposure time, hardness of the circuits and their components, etc.

X-Ray Timing Explorer (XTE) is a GenSAA-based (Generic Spacecraft Analyst Assistant, a superset of Clips)  expert system provided by NASA Goddard. It is a health and safety monitoring rule base, checking the various onboard subsystems on the satellite. Our analysis on this system exposed serious rule naming errors in the system. Detection of such manual errors statically, can provide us with immense confidence on the operational characteristics of the system. Secondly, we were able to identify clusters of rules, which could be suitable candidates for reusability across NASA missions. We found, through the MVP-CA tool, clusters that provided limit checks and status checks on certain telemetry measurands, rules that did background monitoring for telemetry data quality across the different subsystems and also rules that switched the modes of detectors depending on different telemetry values. There can be considerable cost savings to NASA to semi-automatically identify such functional clusters as reusable components for newer NASA missions.

Unexpected Events System (UES), the third expert system that we examined, was built as part of NASA's FUSE (Far Ultraviolet Spectroscopic Explorer) project, by Interface and Control Systems. It is a hybrid system built in SCL (Spacecraft and Command Language), and through its use of scripts and rules, it allows for various operational logics to be defined on the various sensors and subsystems. The UES system performs bit checks on a data stream through digital and analog mnemonics. It has an autonomous system architecture, encompassing both flight and ground operations, as it provides visibility for the state of each on-board sensor or system. The system has over 1000 mnemonics and pseudo-mnemonics that tend to make the system very prone to errors. In the UES system, MVP-CA tool was able to detect several conflicts and anomalies in various pairs of rules. This could lead to very ambiguous behavior of the system at run time. For such a large system, we found that MVP-CA tool was a very valuable tool to "chunk" the system into meaningful units. In particular, when a rule cluster formed around a particular mnemonic, all the values critical for setting the associated pseudo mnemonics could be checked easily. In this way the system could be sliced and diced in various meaningful ways for analysis.

Publications

1. M. Mehrotra, S. Alvarado and R. Wainwright. Laying a Foundation for Reusability of Knowledge Bases in Spacecraft Ground Systems. In Proceedings of the Ground System Architecture Workshop, GSAW99. The Aerospace Corporation. El Segundo, CA. http://sunset.usc.edu/GSAW/1999.
   
   
2. M. Mehrotra, S. Alvarado and R. Wainwright. Laying a Foundation for Software Engineering of Knowledge Bases in Spacecraft Ground Systems. Proceedings of the FLAIRS-99 Conference, to be held in Florida May1-5, 1999.
   
   
3. M. Mehrotra. Exploring Reusability Issues in Telemetry Knowledge Bases. 2nd International Conference on Information Fusion, July 6-8 1999, Silicon Valley, CA.
   
   
4. M. Mehrotra. Multi-ViewPoint Clustering Analysis (MVP-CA) Technology for Mission Rule Set Development and Case-Based Retrieval. Technical Report No. AFRL-VS-TR-1999-1029, Air Force Research Laboratory, Kirtland Air Force Base, NM 87117-5776, April 1999.