A Feature Selection Model to Filter Periodic Variable Stars with Data-sensitive Light-variable Characteristics

At present, autonomous management and operation of spacecraft are the main direction and objective of the development of space technology to lighten the burden of ground measurement and control, reduce the cost of operation and management, and expand the application scope of spacecraft. To select the suitable periodic variable stars with a certain quantity at the given conditions of spacecraft, we study the autonomous navigation method of optical variable spacecraft and propose a feature selection model to filter periodic variable stars with light-variable characteristics. It mainly focuses on the learning processes of the pulsating optical variable light variation star clock model, the high precision pulsating optical variable autonomous navigation algorithm and the optical variable light variation characteristic mechanism with the measurement method. From experiments, the sample of the periodic variable star is selected, forms a database of 132 initial candidate samples and 16 navigation sample stars. So, time measurement can be conducted to take advantage of the nature of periodic variable stars that take days as its cycle, ground-based observation and ground-based application can be conducted with the wide spectrum of periodic variable star observation. It can meet the requirements of spacecraft.