Using Conjunction Analysis Methods for Manoeuvre Detection - Application to Optical Observations

Abstract

Detecting manoeuvres of satellites in a database is an issue to keep those objects in the database instead of having multiple instances. Those multiple instances may occur because manoeuvres lead to difficulties in the object identification and orbit determination process. Efficient manoeuvre detection helps reducing the number of duplicate objects by connecting orbits after a manoeuvre to those before. This work is based on an earlier study where methods traditionally used for collision probability estimation were used to identify manoeuvres. It was possible to calculate a "collision probability" of two osculating element sets, representing two pseudo-objects, and therefore to decide whether a manoeuvre took place. Furthermore, by scanning the interval between both osculating epochs, e.g. calculating one collision probability for each time step, it was possible to estimate the manoeuvre epoch, too. In the transition to optical observations, the distribution of observations of the satellites of interest to determine orbits of sufficient quality is an open issue. One has to find a compromise between dense observation distribution and not focussing solely on that topic. Otherwise, the telescop may be blocked just for that one project, without being intended to. In this study, we compare the former observation strategy to the upcoming one. The satellites are identical to those of the previous study, being Meteosat-8, -9, -10, and -11 operated by EUMETSAT. The optical observations are planned for the telescope network SMARTnet(TM). Due to the geostationary orbits of the satellites, the only affected telescopes of the network are in Sutherland (South Africa) and Zimmerwald (Switzerland), respectively.