DEVELOPMENT OF AN INITIAL TAXONOMY AND CLASSIFICATION SCHEME FOR ARTIFICIAL SPACE OBJECTS

摘要

As space gets more and more populated, a classificationrnscheme based upon scientific taxonomy is needed tornproperly identify, group, and discriminate space objects.rnUsing artificial space object taxonomy also allows forrnscientific understanding of the nature of the space objectrnpopulation and the processes, natural or not, that drivernchanges of an artificial space object class from one tornanother.rnIn a first step, an ancestral-dynamic hierarchicalrntree based on a priori knowledge is established, motivatedrnby taxonomy schemes used in biology. In arnsecond step, available orbital element data has beenrnclustered. Therefore, a normalization of a reducedrnorbital element space has been established to provide arnweighting of the input values. The clustering in the fiverndimensional normalized parameter space is divided inrntwo sub-steps. In a first sub-step, a pre-clustering in arnmodified cluster-feature tree has been applied, to initiallyrngroup the objects and reduce the sheer number of singlernentities, which need to be clustered. In a second sub-step,rna Euclidean minimal tree algorithm has been applied, torndetermine arbitrarily shaped clusters. The clusters alsornallow determination of a passive hazard value for thernsingle clusters, making use of their closest neighbors inrnthe minimal tree and the radar cross section of the clusterrnin question.