Trend Analysis for Spacecraft Systems Using Multimodal Reasoning

Charisse Sary, Chariya Peterson, John Rowe, Troy Ames, Karl Mueller, Walt Truszkowski, and Nigel Ziyad

This paper describes some current work at the National Aeronautics and Space Administration (NASA) Goddard Space Flight Center’s Advanced Architectures and Automation Branch. Trend analysis refers to the process of examining data from a physical system, developing a mathematical model, analyzing the derived information to formulate an evaluation on the condition of the system, and determining if dangerous trends can be detected. If a trend is detected, corrective or preventive actions are pursued. Our goal is to better understand how to effectively use rulebased, case-based and model-based reasoning, together to realize a more rigorous and automated trend analysis capability. To reach this goal, we plan to develop an automated system to analyze and predict trends, and diagnose spacecraft status telemetry data. This paper describes a concept, architecture and current work in developing a prototype system, called the Automated Model-Based Trend Analysis System (AMTAS). This system uses multimodal reasoning to perform diagnosis and trend analysis. Model-based reasoning is the primary reasoning component which is augmented with other forms of reasoning including rule-based reasoning and case-based reasoning. This prototype may serve as a basis for a full system implementation at a later time if successful. We are in the process of implementing the prototype system using MATLAB.


This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.