The Integrations of an Expert System into a Real-Time Application

Abstract

The proposed NASA space station will depend heavily upon state-of-the-art technologies. One area of active research is in the automation of the space station through the use of Artificial Intelligence (AI). One practical use of AI is in the area of expert systems. Some expert systems are being proposed to monitor the station's power distribution and management. Due to the nature of this problem, several unique constraints upon an expert system have been imposed including: mass, volume, cost, and speed. This research addresses these constraints and provides a possible solution for power control aboard a NASA space station. In addition, this research also addresses the more fundamental question of achieving real-time performance in an expert system. Significant results were achieved and are documented that provide real-time expert system performance for certain problem domains.