Satellite maneuver detection and estimation using optical measurements

Abstract

The maintenance of a Space surveillance orbit catalogue requires a robust orbit determination process to update the state of each catalogued object. For the case of active-operated satellites, the process of orbit estimation may lead to great state uncertainties if no-maneuver information are available at the moment of determining the orbit. This thesis, firstly, proposes the analysis of four post orbit determination methods for maneuver detection. In particular, two of them are already-existing techniques found in the literature, while the others are innovative and entirely developed during this work. Their detection capabilities have been subsequently tested and compared, revealing that best performances are provided by the newly-implemented Linear Kalman filter for maneuver detection. Secondly, an entire orbit determination process, including the detection of maneuvers, is described. An Extended Kalman filter is in fact used to estimate the orbit of the telecommunication satellite ASTRA-1KR from optical measurements and to infer consideration related to the observations geometry. Subsequently the problem of refining the detected maneuver time and of estimating the maneuver components is addressed. The work is concluded with the proposal of a task-flow, illustrating all the actions a space surveillance orbit catalogue should perform for detecting maneuvers.