An Adaptive Satellite Orbit Estimation Scheme Using an Analytical Approach and a Dynamical Model Compensation Technique

摘要

An Earth satellite orbit estimation procedure for real time applications is presented. A perturbed two body physical model is considered for the satellite equations of motion in the mathematical model of the estimator. Using the KS transformation, an analytical solution is obtained in the correspondent parametric transformed space. The perturbations in the Keplarian motion are treated, in the parametric space, as unmodeled accelerations. A dynamical model compensation technique approximates each component of these accelerations by a first order Gauss-Markov process. These accelerations are estimated along with the satellite motion by the definition of an extended state, using the extended Kalman filter. During the estimation process, the observations nonlinearities and the state noise covariance matrix are adaptively determined to prevent filter divergence. Results of numerical tests under simulated conditions are presented and discussed.