A monitoring system with tolerance for real-time data problems

Summary form only given. To overcome the problems related to real-time data, the authors have developed a set of methods which allow a monitoring system to tolerate bad data, or to automatically recover if false conclusions are made based on bad data. DESSY (DEcision Support System) is a rule-based system which monitors Space Shuttle telemetry and makes intelligent evaluations of system hardware through assessments of telemetry data patterns and transitions. A set of methods is proposed to handle data problems that are based on strategies used by the human experts. They include rule disabling, use of context and expectations to make assessments that tolerate some bad data, and graceful recovery through system correction when reliable data return. Self correction is an important feature for the real-time expert system. DESSY has been implemented in a commercial real-time expert system development environment. The system currently monitors approximately 80 pieces of Shuttle telemetry. The robustness of DESSY makes it capable of lengthy periods of uninterrupted use in real-time operations