Seminars

Abstract: As the number of objects in Earth orbit increases, so does the need to improve the level of automation of Space Surveillance and Tracking (SST) systems. One of the main problems to be addressed in this regard are orbital maneuvers performed by non-cooperative spacecraft, which can lead to a loss of object custody or significantly biased state estimates. Different approaches have been proposed to solve this problem, most of them relying on some optimality criteria or machine learning methods based on historical patterns. It can be argued that depending on the observability conditions, determining an optimal transfer subject to certain observed quantities may lead to biased estimation and a posterior filter divergence. However, in the absence of prior maneuver knowledge, fuel efficiency and energy considerations are the only reasonable drivers for post-maneuver state estimation. Therefore, we propose to develop a framework that allows for both optimal and heuristic based estimation in an online fashion. A key aspect of maneuver detection is the computation of the prior density, which is typically associated with the so-called reachable set. A tractable solution to this problem is sought by the definition of computationally efficient control distance metrics, which enable a fast evaluation of the distance between two orbits subject to certain time of flight and propulsion type. The latter allows for sampling-based approaches to be used in approximating the posterior density of the state in the event of unknown maneuvers, thus increasing the accuracy and consistency of the main tracking filter.

Bio: Guillermo Escribano is a PhD candidate in Aerospace Engineering at Universidad Carlos III de Madrid, where he also received a BSc in Aerospace Engineering and an MSc in Aeronautical Engineering. His PhD research is in collaboration with the European Space Agency (ESA) and GMV, and focuses on maneuvering target tracking applied to Space Surveillance and Tracking. Other lines of research include uncertainty propagation, data association and collision risk analysis.