Research On Algorithms For Real-time Orbit Determination Using On-board GPS

Algorithms of on-board GPS real-time single point positioning and dynamic orbit determination using positioning solutions are mainly studied.First, several practical problems in GPS positioning are investigated. As to the poor performance of Klobuchar ionospheric correction model, a new method using a constant scale factor is presented; GPS satellites with negative elevation angles are found in observations, therefore, they should be deleted at positioning to improve observation geometry and precision. With the respect to effects of initial values on positioning, iterative computations should be executed after getting direct solutions from GPS equations when positioning intervals increase, especially longer than 5 seconds.An improved positioning method of piecewise code smoothing is introduced, which freezes the weighting factor in the process of computation to keep the weight of current observations. The smoothing results are good when the frozen weighting factor is between 10 and 15.An ionosphere-free positioning method is brought forward, which is derived from DRVID model to eliminate the effects of ionosphere. The results show that it improves positioning accuracy a lot, about 2 to 5 meters.An algorithm of dynamic orbit determination using positioning solutions is studied, and the positioning trajectory is smoothed out by the dynamic model, which reduces random error and improves orbit accuracy greatly. The algorithm uses an improved extended Kalman filter with a least square estimator parallel to it, which monitors the filter's state and restrains divergence without increasing the burden of computation. The algorithm and related influencing factors, such as gravitational model, atmospheric drag, sample intervals and filter parameters, are deeply analyzed, and some important conclusions are deduced.To sum up, tested and validated by data sets of Shenzhou IV unmanned spacecraft and precise orbit ephemeris, the algorithms presented in the thesis can improve orbit accuracy and reduce random noise considerablely. Derived conclusions have important value for reference, and lay the firm technological foundation for future applications.