Error assessment of an autonomous real-time precision orbit determination program for a low-earth-orbit satellite using GPS observation data

An autonomous, real-time, precision orbit determination (ARTPOD) program for low-Earth-orbit (LEO) satellites using the Global Positioning System (GPS) is planned and developed. Numerical simulations are used to assess the anticipated accuracy of LEO satellite tracking using GPS observation data. The GPS observation data for a particular LEO satellite, CHAMP (Challenging Mini-satellite Payload), is processed with the ARTPOD program and results are compared with those from a current reference level Precision Orbit Determination (POD) program. A set of gravitational and non-gravitational force model and measurement errors is applied to provide anticipated levels of orbit error in the estimation process. Non-differenced LEO-GPS pseudo measurements are used to improve the GPS data model and the Earth geopotential model. The latter is the dominant error source for LEO Precision Orbit Determination (POD). Several representations of the gravitation model are evaluated to find the optimum model based on two criteria: performance and size of program. A batch filter and an extended Kalman filter are used to provide statistical best state estimates. The design of ARTPOD is evaluated to determine its suitability for autonomous, real-time operation. Results presented show that the batch filter and extended Kalman filter are successfully implemented in ARTPOD. These two filters converged in the worst case scenario based on initial state. The final orbit was estimated within anticipated errors with nominal initial state. Various combinations of initial state and estimate state vector are tested to improve the ARTPOD design. The results by four different methods of obit improvement is shown.