A Comparison of the Extended Kalman Filter and Weighted Least Squares in Early-Orbit Determination
AIR FORCE INST OF TECH WRIGHT-PATTERSON AFB OH SCHOOL OF ENGINEERING
Pagination or Media Count:
The problem of estimating the orbit parameters from early-orbit observations of an earth satellite is used to compare the accuracy and application of the extended Kalman filter and the classifical filtering method of weighted least squares. To obtain an absolute comparison, a treu two-body, drag-free Keplerian orbit is simulated, observations are computed and contaminated with noise, and the orbit parameters estimated by each filter are compared. The accuracy of the two filters was compared using the same set of observations to determine the effect of observation truncation and initial conditions on the results. Based on this study it was concluded that the weighted least squares filter and the extended Kalman filter yield about the same accuracy in the early-orbit determination problem.
- Spacecraft Trajectories and Reentry