R. Harting, M. Miller

Successful monitoring of a complex system, similar to quality control tasks, requires the accurate analysis of numerous measures which reflect the operational status of the process. Due to the complex, non-linear interaction of these measures, their effect on the process status are oftentimes not fully understood. Neural networks solve such high-dimensional non-linear problems by supporting a data-driven synthesis of a valid functional mapping from a set of system status measures to system state. With this mapping, a significant reduction in both costs and fall-out periods can be attained through early problem recognition--before costly critical situations occur. Also, analytical investigation helps the user understand system interactions (e.g., identifying comparative relevance of various influencing factors).

DGLR Symposium, Stuttgart, 1995

Conference Paper

englisch

17,0 x 24 cm, 11 Seiten

DGLR-Bericht, 1995, 1995-03, Aircraft Integrated Monitoring Systems; S.663-673; 1995; Deutsche Gesellschaft für Luft- und Raumfahrt - Lilienthal-Oberth e.V., Bonn

NA

monitoring, quality control, neural nets

Bestellbar

1995