Effects of star detection errors on attitude determination

In this paper, robustness of star identification and attitude estimation methods that are operating in Lost-in-Space (LIS) mode, under harsh noise conditions of the near space, are analyzed. Despite star extraction, identification and attitude estimation methods that are proposed as solutions for subproblems, there is a lack of study that investigates the effects of errors in the initial stages on the overall performance. This work addresses this issue by testing two prominent identification methods in tandem with two canonical attitude determination methods. The experiments are performed on the publicly available Yale Bright Star Catalog and the results are analyzed with ground truth-based evaluations.