Processing of Space Surveillance Observations

This thesis covers the processing of different types of space surveillance measurements. The emphasis is on the initial build-up of a catalogue. Most experiments focus on the identification of objects from pairs of observations by testing whether these two observations could originate from the same object. The advantage of this approach is that an orbit from two passes is usually more precise than that from a single pass and thus the chance of associating subsequent measurements with this newly derived orbit is higher. The main contribution is the introduction of a method for perturbed initial orbit determination and observation correlation from two radar tracklets. The perturbed initial orbit determination provides the solutions for all possible numbers of revolutions from two positions under consideration of the secular J2 -perturbations, which causes for example the rotation of the orbital plane. The perturbations are considered analytically in an iterative process. From the different numbers of revolutions, the most probable solution can either be selected via the minimum Mahalanobis distance, using the range-rate as a remaining observable, or after an additional post-processing. The robustness and applicability of the method is shown using real radar measurements, which includes a large percentage of successful initial orbit determinations for tracklets which are more than 20 days apart. This method is also extended to the combination of optical and radar measurements for objects in Highly Elliptical Orbits. Concerning the simulation of a cold-start of a catalogue, two different processing strategies are introduced. One is adding a least squares orbit determination using a pair of possibly correlated tracklets, while the other approach collects data over a longer time span and generates a graph network from which clusters of tracklets are derived as new candidate objects. Further analysis focuses on aspects of operational processing of space surveillance measurements. This includes the automated decision making process for the monitoring of the quality of existing catalogue objects and the design of an autonomous processing pipeline for optical measurements at the Zimmerwald observatory. A potential space mission to observe the Geostationary Orbit is also analysed with regard to its performance during the catalogue build-up