Autonomous State Determination for an Earth Orbiting Satellite Using Horizon and Star Sensors

摘要

An autonomous state determination system is developed for an Earth orbiting satellite using horizon sensors and star trackers. The horizon sensors detect the Earth and establish the local vertical reference, while either one or two star trackers make angular sightings of known stars. These sightings constitute the observation function used in special perturbations differential correction techniques to produce the state estimate. Two modes are used, batch and sequential. The batch method uses nonlinear least squares estimation while the sequential mode uses Bayes estimation. Both modes rely upon residuals to form state corrections. Truth model generated sightings are used to form residuals. Three separate estimation versions are tested: nonlinear least squares, Bayes estimation with fixed data, and Bayes estimation with changing data. Each of these three routines exist in one and two star tracker versions. Eleven test cases are run with the different estimator routines and indicate better than anticipated results. Problem areas include estimator dependence upon accurate initial estimate, sensitivity to specific perturbed orbital elements, and varying minimum number of data points for Bayes estimator.