Sensor Data Qualification for Autonomous Operation of Space Systems

William A. Maul, Kevin J. Melcher, Amy K. Chicatelli, T. Shane Showers

NASA's new Exploration initiative for both robotic and manned missions will require higher levels of reliability, autonomy and re-configuration capability to make the missions safe, successful and affordable. Future systems will require diagnostic reasoning to assess the health of the system in order to maintain the system's functionality. The diagnostic reasoning and assessment will involve data qualification, fault detection, fault isolation and remediation control. A team of researchers at the NASA Glenn Research Center is currently working on a Sensor Data Qualification (SDQ) system that will support these critical evaluation processes, for both automated and human-in-the-loop applications. Data qualification is required as a first step so that critical safety and operational decisions are based on "good" data. The SDQ system would monitor a network of related sensors to determine the health of individual sensors within that network. Various diagnostic systems such as the Caution and Warning System would then use the sensor health information with confidence. The proposed SDQ technology will be demonstrated on a variety of subsystems that are relevant to NASA's Exploration systems, which currently include an electrical power system (EPS) and a cryogenic fluid management system. The focus of this paper is the development and demonstration of a SDQ application for a prototype power distribution unit that is representative of a Crew Exploration Vehicle electrical power system; this provides a unique and relevant environment in which to demonstrate the feasibility of the SDQ technology.

Subjects: 1.5 Diagnosis; 3.4 Probabilistic Reasoning

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.