摘要：AbstractPerformance of an EKF-based Orbit Determination (OD) algorithm using Global Positioning System (GPS) measurements is examined under erroneous observation conditions. A generalized formulation of a dynamic orbit model is designed to predict the satellite states in an Earth-Centered Earth-Fixed (ECEF) frame. A suitable orbit interpolation method is then chosen for interpolating the GPS satellite position at the measurement time tag before finding the geometric range between the GPS and the LEO satellites. Thereafter, the commonly used Extended Kalman Filter (EKF) algorithm is implemented for the nonlinear OD problem. The proposed algorithm is tested for the precision in OD of two LEO satellites, namely GRACE and Megha-Tropiques based on the available raw GPS measurements in the Receiver Independent Exchange (RINEX) observation files. For the GRACE satellite, the efficacy of the present OD algorithm is validated by comparing the estimated and publicly available precise orbit information. The Megha-Tropiques satellite, however, does not have precise orbit information. Furthermore, it is observed that the RINEX file has several erroneous observations as detected by the Fault Detection and Exclusion (FDE) algorithm in-built in the OD algorithm. A comparison of the estimated satellite position with that obtained from the standard software GNSS-Lab Tool (gLAB) shows that the accuracy of the position estimates varies proportionally with the number of erroneous observations of the Megha-Tropiques satellite.
