Simão da Graça Marto, University of Strathclyde; Massimiliano Vasile, University of Strathclyde; Sebastian Diaz Riofrio, University of Strathclyde; Christos Ilioudis, University of Strathclyde; Carmine Clemente, University of Strathclyde
Keywords: Bistatic radar, behaviour analysis, SSA
Abstract:
Since their inception, ground based Space Situational Awareness (SSA) systems, also known as Space Surveillance and Tracking (SST), primarily utilise radar sensors due to their ability to operate in very long ranges and under various atmospheric conditions while also providing very accurate range measurements. Initially used for early missile warning, modern SST radars are designed to monitor targets in Low Earth Orbit (LEO), up to deep space. Having very high power transmissions, in order to improve their efficiency, radar systems can also operate in tandem with nearby radio frequency (RF) telescopes forming what is known as bistatic configuration. In this configuration, the reflected signal is received not only by the primary emitter station, but also by distant RF telescopes.
A prime example of such bistatic system is the Tracking and Imaging Radar (TIRA) located at Fraunhofer Institute for High Frequency Physics and Radar Techniques (FHR), Germany and the Effelsberg radio telescope which when paired can improve the minimum target size from 2cm at 1000 km down to 1cm due to the higher sensitivity of Effelsberg, a 100m diameter dish radio telescope from the Max-Planck institute. While technically bistatic, due to the very high target altitudes the experienced bistatic angles are generally small and the perceived radar cross section (RCS) of the target will be very similar to that of a monostatic radar. Recently, the use of long baseline bistatic radar systems for SST was proposed in the SST community. The difference with existing bistatic systems is that the RF telescopes are remotely located from the radar allowing larger bistatic angles and essentially viewing the target from different aspect angles.
Specifically, capturing the target reflection in different bistatic angles can result in higher RCS than the monostatic depending on its shape, while integrating multiple measurements results in a higher signal to noise ratio (SNR). Moreover, combining location and velocity estimates of a target from distributed sensors can significantly improve the parameter estimation accuracy.
While preliminary analysis on captures of GEO satellites has shown that spaceborne targets can be detected in these bistatic configurations, covered research lacks from a more extensive investigation of the bespoke processing framework and performance analysis aimed at fully unlocking the benefits of such a system.
The use of tracking data from radar systems allows reconstructing the motion of space objects beyond the simple orbit determination. If it is assumed that the object in view has a behaviour dictated by an unknown part of the dynamics, it is possible to use modern machine learning techniques to reconstruct the missing part of the dynamics and recognise patterns or intentions. This process is called behaviour analysis.
In this work we will assess the benefits and trade-offs of multiple behaviour analysis methods proposed in the literature, which we apply to this use-case of spacecraft being observed by multistatic radar. Optimal control based analyses proposed [Serra et al. 2022; Lubey and Scheeres 2014], as well as statistical approaches are some examples of behaviour analysis techniques to be considered. In general, they provide a metric that measures the likelihood that a manoeuvre was performed, and/or a mathematical description of this manoeuvre.
We will also investigate to what extent the increased data quality coming from bistatic radar observations improves the accuracy of the results of the behaviour analysis process when compared to the monostatic case. We will consider in particular two possible scenarios: an orbit repositioning to acquire new targets on ground and an orbit re-positioning to shadow another satellite. These two case studies will be used to assess the utility of combining direct observation data, including the time evolution of the multistatic radar data, together with the derived reconstruction of missing parts of the dynamics to infer the possible long-term behaviour of an object.
In summary, this work aims to to obtain concrete information on the benefits of using a bistatic setup compared to a monostatic one for the purposes of behavioural analysis, when applied to detecting targets in LEO altitudes in realistic scenarios.
[Serra et al. 2022] – Serra, Romain, Carlos Yanez, and Carolin Frueh. “Tracklet-to-orbit association for maneuvering space objects using optimal control theory.” Acta Astronautica 181 (2021): 271-281.
[Lubey and Scheeres 2014] – Lubey, Daniel P., and Daniel J. Scheeres. “Combined optimal control and state estimation for the purposes of maneuver detection and reconstruction.” 2014 American Control Conference. IEEE, 2014.
[Vasile 2019] – Vasile, Massimiliano. “Polynomial representation of model uncertainty in dynamical systems.” Advances in Evolutionary and Deterministic Methods for Design, Optimization and Control in Engineering and Sciences. Springer, Cham, 2019. 419-432.
Date of Conference: September 27-20, 2022
Track: SSA/SDA