A Study On The Precise Orbit Determination Of Geostationary Satellites

This dissertation mainly focuses on Precision Orbit Determination (POD) and fast post-maneuver orbit recovery of GEO satellites with the accuracy of 10-1 meter. Two commonly used yet powerful tools, variance-covariance matrix analysis and Monte-Carlo simulation, are used to investigate orbital errors and precision comparison. The former is specifically utilized to analyze the data set of ground-based range and VLBI/CEI phase with a simplified analytic algorithm developed in this paper, while the latter is applied to inter-satellites range and range rate (RARA). Finally the fast orbit recovery within several hours after station-keeping maneuvers of GEO is discussed. The main works and contributions are summarized as follows:1. The estimate and covariance formula of self-calibration POD are derived by using QR factorization. It is proven that, compared with POD without self-calibration, self-calibration POD has the merits of calibrating some systematic errors in the measurement model or force model and improving the accuracy of orbit determination simultaneously. With the constraints of geostationary orbit and high altitude, the effects onto POD resulting from observational errors are investigated. The covariance analysis formula and its approximate algorithm of orbit determination are established.2. Both advantages and weakness are analysed among several groung-based GEOs'tracking techniques. The formal behavior of systematic errors in classical range-based POD is pointed out, namely, at the first-order approximation, constant or piecewise constant range biases will be equivalently modeled for the representative of the summation of all kinds of error sources in the measurement residuals (O-C), due to the difficulty of distinguishing one error from another. It is presented that mini-scale network of two sets CEIs with large span of longitude will be a novel tracking and orbit determination solution, along with one set CEI augmented with few quantity of SLR data.The feasibility and potential application are discussed for VLBI and CEI in the 10m POD of GEOs' solutions. Results show that systematic errors (namely, range biases or interferometric phase biases) are the dominant factors to achieve 10m-level POD. They may be 8-20 times magnified to the transversal position uncertainty according to different number and distribution of the tracking sites. With the arc length enlarged to 24 hours, the uncertainty of cross-track positional component can be decreased to no more than 10 meters, while along-track position can not be improved apparently.