Justification of the choice of the method and criterion of clustering for intelligent analysis in flight control spacecraft

Abstract

The article examines the methods of intelligent analysis of telemetric information of spacecraft. The current state and main shortcomings of the control process during the spacecraft flight control are briefly given. It is proposed to eliminate the shortcomings by introducing intellectualization procedures in terms of telemetric information analysis. Based on the methods of cluster data analysis, a method is proposed for automatically determining the moment of occurrence of anomalies in the state of a spacecraft, which are precursors of off-nominal situations. A schematic diagram of the operation of an intelligent control system based on the use of the method of cluster analysis of the spacecraft telemetric information is presented. The conditions for choosing the method and criterion for clustering are substantiated, taking into account the goals pursued in solving control problems during flight control of the spacecraft. A mathematical description of the clustering methods and criteria selected for further practical testing is given. To test the proposed method of analysis for various methods and criteria of clustering, calculations were performed using archived telemetric information. From the point of view of the time of early detection of the anomaly in the state for a separate component of the spacecraft, the choice of the method and criterion of clustering is made for further research and experimental work.