Development of a second-order sliding innovation filter for an aerospace system

Abstract

The sliding innovation filter is a new type of predictor-corrector estimation method. The strategy is used to estimate relevant states of interests and has been found to be robust to modeling uncertainties and disturbances. In this paper, a second-order formulation of the sliding innovation filter is presented to improve its estimation performance in terms of accuracy while maintaining robustness. The strategy is applied to an aerospace system that has been designed for experimentation. The results are compared with the well-known Kalman filter, and future work is considered.