Forecasting time series of spacecraft telemetry based on ensembles of neural networks
Abstract
The article describes a two-level model of neural network ensembles for predicting multidimensional telemetry time series of spacecraft subsystems. The developed model is implemented in the system for identifying their state by telemetry data for the ground command-measuring complex.
About the Authors
A. Doudkin
Объединенный институт проблем информатики НАН Беларуси
Belarus
Belarus
Y. Marushko
Объединенный институт проблем информатики НАН Беларуси
Belarus
Belarus
V. Ganchenko
Объединенный институт проблем информатики НАН Беларуси
Belarus
Belarus
References
1. Short-Term Load and Wind Power Forecasting Using Neural Network-Based Prediction Intervals / Hao Quan [et al.] // IEEE Transactions on Neural Networks and Learning Systems. – Short-Term Load and Wind Power Forecasting Using Neural Network-Based Prediction Intervals. – 2013. Vol. 25. Iss. 2. ISSN: 2162–237X. P. 303–315.
2. Lysayak A. S. Prediction of multidimensional time series / A. S. Lysayak, B. Y. Ryabko // Vestnik SibGUTI. 2014. №4. P. 75–88.
3. Comparison of the ARMA, ARIMA, and the autoregressive artificial neural network models in forecasting the monthly inflow of Dez dam reservoir / M. Valipour [et al.] // Journal of Hydrology 476 (2013). P. 433–441.
4. Khachumov V. M. Review of Standards and the concept of monitoring, control and diagnostics of the spacecraft tools building / V. M. Khachumov [et al.] // Software Systems: Theory and Applications. №3(26). Vol. 6. 2015. P. 21–43.
5. Emelyanov Yu.G.: Neural orientation angles and distance of the spacecraft sensor control system / Yu. G. Emelyanov, K. A. Konstantinov, S. V. Pogodin, etc. // Software Systems: Theory and Applications. 2010. №1(1). P. 45–59.
6. Neural network ensemble operators for time series forecasting / Nikolaos Kourentzes [et al.] // Expert Systems with Applications. July 2014. Vol.41. Iss. 9. ISSN: 0957–4174. P. 4235–4244.
Review
For citations:
Doudkin A., Marushko Y., Ganchenko V. Forecasting time series of spacecraft telemetry based on ensembles of neural networks. Science and Innovations. 2021;(5):16-22. (In Russ.)
Views: 84
Email this article 