Reducing variability in short term orbital lifetime prediction

Within the last year three major re-entries occurred. The satellites UARS, ROSAT and Phobos-Grunt entered Earth’s atmosphere with fragments reaching the surface. Due to a number of uncertainties in propagating an object’s trajectory the exact place and time of a satellite’s re-entry is hard to determine. Major influences when predicting the re-entry time are the changing precision of the available orbital data, the satellite’s ballistic coefficient, the activity of the sun which influences the Earth’s atmosphere and the underlying quality of the atmospheric model. In this paper a method is presented which can reduce the variability in short-term orbital lifetime prediction induced by fluctuating orbital data accuracies. A re-entry campaign is used as a reference for this purpose. For a window of a few weeks before the re-entry the position data of a synthetic object is disturbed considering different degrees of orbital data errors. As a result different predictions will exist for the generated position data of a given day. Using a regression algorithm on the available data an average position is obtained, which is then used for the orbital lifetime prediction. The effect of this measure is a more consistent prediction of the orbital lifetime. The paper concludes with the comparison of the generated re-entry windows in various test cases for the original and the averaged data.