Abstract
A large number of nonstationary, nonlinear and periodic telemetry data have been collected to support the prognostics and health management of satellite which are not suitable for traditional time series prediction models. This paper proposes a PSRT-RVM model based on the joint optimization of quantum genetic algorithm (QGA) for on-orbit prediction. The method first separates the trend terms based on the time series decomposition theory to stabilize the remaining terms. The PSRT-RVM model optimized by QGA is used to predict the residual term while the trend term is fitted with a polynomial. Finally, the two predicted values were recombined to get the predicted results. The feasibility and effectiveness of this method are verified by the comparative experiment of the periodic telemetry parameters of a satellite in China.

This publication has 12 references indexed in Scilit: