Among the priority tasks in the creation of new generation rocket and space technology, a special place is held by the task of maintaining and extending the life cycle of spacecraft. For this purpose, automated systems are being created that monitor and predict the state of onboard subsystems of spacecraft and contribute to increasing their durability. The increasing requirements for the characteristics of space systems inevitably leads to a revision of the used control technologies and the need to create a scientific and technical reserve in the form of approaches, methods and technologies for constructing promising competitive space technology. To increase the reliability and extend the service life of spacecraft under conditions of aging, information loss and natural noise, technologies (algorithms and methods) are needed for constructing integrated information support for solving the problems of data recovery, monitoring and predicting the states of spacecraft subsystems; applied object-oriented artificial intelligence systems: artificial neural networks and neurocomputers for control, diagnostics and decision support. In this paper, methods of neural network monitoring of the state of onboard subsystems of small spacecraft are proposed and investigated. In this work, various neural networks were developed and trained to solve a number of urgent problems of monitoring the operation of spacecraft subsystems. Methods for high-precision reconstruction, forecasting of time series and monitoring the state of spacecraft subsystems using telemetry data were proposed.
DOI: 10.1063/5.0099226
В базе Института программных систем имени А. К. Айламазяна РАН: http://skif.pereslavl.ru/psi-info/rcms/rcms-publications/2022.html
ResearchGate: https://www.researchgate.net/publication/362732218_Neural_network_technologies_for_control_of_subsystems_of_small_spacecraft
Shishkin O. G., Khachumov V. M. Neural Network Technologies for Control of Subsystems of Small Spacecraft // Proceedings of the International Conference on Engineering Research 2021 (ICER 2021), pp. 1-8.