Optimization of Observation Strategy to Improve Re-Entry Prediction of Objects in HEO

During the last decade the number of space debris moving on high elliptical orbit (HEO) has grown fast. Many of these resident space objects (RSO) consist of medium and large spent upper stages of launch vehicles, whose atmosphere re-entry might violate on-ground casualty risk constraints. Increasing the accuracy of re-entry predictions for this class of RSO is therefore a key issue to limit the hazards on the Earth assets. Traditional computational methods are mainly based on the exploitation of Two Line Elements (TLEs), provided by the United States Strategic Command (USSTRATCOM) and currently the only public data source available for these kind of analyses. TLE data however, are characterized by low accuracies, and in general come without any uncertainty information, thus limiting the achievable precision of the re-entry estimates. Better results on the other hand, can be obtained through the exploitation of observational data provided by one or more Earth sensors. Despite the benefits, this approach introduces a whole new set of complexities, mainly related with the design of proper observation campaigns. This paper presents a method based on evolutionary algorithms, for the optimization of observation strategies. The effectiveness of the proposed approach is demonstrated through dedicated examples, in which re-entry predictions, attainable with existing and ideal sensor architectures, are compared with corresponding results derived from TLE data.