Space object data association using spatial pattern recognition approaches

Closely-spaced objects, especially debris objects, create a setting that is very similar to a multi-target environment in a tracking problem. This environment engenders a major data association problem in the field of space situational awareness. To address this problem, an approach that couples gating methods for data association along with a star pattern recognition algorithm, called the planar triangular method, is developed. The planar triangular method has been shown to work effectively for spacecraft attitude determination using star trackers by comparing stars in field-of-view to those present in the catalog. This approach is further enhanced here for association of resident space objects. The planar triangle approach is further enhanced to associate closely spaced objects by incorporating a classical validation gate-based algorithm. The work in this paper shows the effectiveness of combining traditional data association methods with an existing planar triangle pattern recognition algorithm for space object association. Results indicate that the traditional gating algorithm significantly improves the planar triangular method's accuracy for space object association of closely-spaced clutters in highly uncertain environments.