The paper discusses an approach to model epistemic uncertainty in Conjunction Data Messages (CDMs) and classify conjunction events based on the confidence in the collision probability. The proposed approach uses the Dempster-Shafer Theory (DSt) of evidence and assumes that the observed CDMs are from a family of unknown distributions. The Dvoretzky Kiefer Wolfowitz (DKW) inequality is used to construct robust bounds on these unknown distributions. The DSt structure encapsulates the uncertainty in the CDMs and allows the computation of belief and plausibility in the realisation of a given collision probability. The methodology is tested on real events and compared against existing practices in the European and French Space Agencies.
Publication date: 28 Jan 2024
Project Page: arXiv:2402.00060v1
Paper: https://arxiv.org/pdf/2402.00060
