Optimal Filtering for LISA with Effective System Models

Abstract

Two effective dynamic models have been designed and studied in this chapter. These effective dynamic models employ smaller state vector and simplified dynamic functions, hence they are numerically more efficient and more robust. The degeneracy in the previous Kalman filter model can be eliminated by these effective models. The posterior measurements have been used via a RTS smoother and a modified iterative smoother to improve the estimates of the algorithms. Simulation shows that the posterior measurements can further reduce the estimation errors significantly.