The trend prediction for spacecraft state based on wavelet analysis and time series method

Abstract

Based on a large number of downlink telemetry data during the spacecraft on-orbit operation, the characteristic of spacecraft state change is obtained. It is of great significance to realize the safe and reliable spacecraft operation management. In order to achieve the accurate trend prediction for a spacecraft, a hybrid prediction algorithm using wavelet analysis and time series method is presented on the basis of mechanism analysis and data characteristics analysis. Firstly, wavelet analysis is introduced to make decomposition and reconstruction calculations for downlink telemetry signals, and non-stationary signal can be converted to multi-layer relatively stable decomposition sequences. Secondly, a prediction model for each decomposition level sequence is established by using the method of time series. Finally, the final prediction results can be obtained by adding the predicted value of each layer. The simulation results show that the combined model not only has higher prediction accuracy, but also have stronger adaptability for different forecast objects. The method can provide evidence for improving the validity and correctness of spacecraft data analysis and fault diagnosis.