?
Алгоритмы и программное обеспечение автоматизации процессов измерений и обработки данных оперативной диагностики ионосферы и ионосферных радиолиний
The article contains the main architectural principles and technical solutions in the development of the software part of the hardware and software complex for the operational diagnostics of the ionosphere and ionospheric radio channels, mainly through remote ground-based radiosonding. A brief description of measurement techniques and measuring equipment is given. The tasks of automating the measurement process itself and data recording have been solved, which allow organizing the work of the complex according to a given schedule without direct human participation. The tasks of extracting information from the obtained experimental data of radiosonding and ionosphere diagnostics have also been solved, including the possibilities for working with the hierarchy of experimental data, including the possibility of degeneration of derived data, the possibility of viewing available data and the possibility of batch processing of large data arrays to obtain and study their statistical characteristics. The organization of processing when collecting statistics of variations of intermode delays, amplitude-frequency characteristics of partial modes of propagation of short waves, as well as classification of these data according to the so-called empirical multipath models are considered; examples of results are given. The prospects and preliminary results of applying the achievements of the machine learning to solve the problems of automating the extraction of information from radiosonding data are shown; the software of marking data for the formation of training samples is described. The plans for further research are the development of quality metrics and the choice of optimal architectures of deep neural networks, as well as the development of markup tools, including instruments of cross-monitoring the results.