The study of spacecraft telemetry data prediction based-on SERTS model

This paper studies a real-time prediction method associated with the state and event of spacecraft in-orbit. The process includes the following steps: 1) Access the information about the parameter and state configuration related to the state and event of spacecraft, and construct the SERTS model (State-Event-Related-Time-Series model) of the changed parameters; 2) Select the parameters used to characterize the status of the system in the timing method to compose the vector, and extract linearly independent vectors; 3) Construct the distance discriminant functions, and identify the classification threshold value of the status according to the dynamic change of the system status; 4) Monitor the spacecraft in-orbit in real-time and predict the trend of the parameter according to the above distance discriminant function and the threshold. Besides, the adaptive threshold selection method during the process of SERTS modeling and predicting has also been debated. Finally, the experimental result shows that the SERTS model can be effectively applied to the predicting and alarming of the telemetry data related to the state and event of spacecraft. The proposed SERTS model and method can provide the efficient tool for the satellite operation and maintenance staff and provide the automated means to improve work efficiency.