Scientific journal
Scientific Review. Technical science
ISSN 2500-0799
ПИ №ФС77-57440

Personal portfolio

APPLICATION OF ARTIFICIAL NEURAL NETWORKS IN SPACECRAFT SUBSYSTEMS CONTROL AND DIAGNOSIS PROBLEMS

Abramov N.S. 1 Zadneprovsky V.F. 1 Talalaev A.A. 1 Fralenko V.P. 1
1 Ailamazyan Program Systems Institute of the Russian Academy of Sciences
In this paper is carried out a review of recent scientific publications in the field of building intelligent systems for spacecraft subsystems control and diagnosis. A special attention is given to issues of artificial neural network technology application that implement intelligent functions by their abilities to learn and generalize information, detect and predict data streams anomalies. They allow solve with high accuracy and completeness the classification problem (recognition) situations, increasing the rate of production decisions and ensuring the reliability of the satellite units. Possible fields of their application: control, diagnostics and prediction of subsystem states, encryption of data transmitted over the communication channels. Two schemes of spacecraft state analysis are presented: with earth component using and directly on board. Addition of neural networks in the telemetry, command and target information processing tools enhances the functionality, reliability and increase an autonomy of both ground and airborne control complexes.