Calum Meredith, Defence Science Technology Laboratory (Dstl); Paul Chote, University of Warwick; Robert Airey, University of Warwick
Keywords: SDA, RPO, Non-Resolved Imaging
Abstract:
The PHANTOM ECHOES 2 experiment (2020-2021) was an international research collaboration involving the defence science and technology agencies of Australia, Canada, New Zealand, the United Kingdom and the United States, and performed under The Technical Cooperation Program. It studied the rendezvous and docking of Mission Extension Vehicle 2 (MEV-2) and Intelsat 10-02 in geostationary Earth orbit (GEO) utilising data from multiple sensors to observe and scrutinise the dynamics and behaviours of the two satellites. This experiment presented the opportunity to gather data on these satellites whilst exhibiting varying levels of proximity at GEO, as well as the chance to observe both satellites separately and to obtain ground truth data on true separation distances.
Maintaining tracks on GEO spacecraft and other closely-spaced objects when they are in close proximity has key importance within Space Domain Awareness for understanding satellite cluster behaviour, conjunction events and the intentions of closely manoeuvring satellites. When two satellites carry out rendezvous and proximity operations, or similar activities, the ability to separate and distinguish between the two objects using electro-optical techniques can be challenging: at a particularly close separation, the two objects appear merged within an image and become unable to be individually resolved. However, if one object displays glinting behaviour (for example, whilst manoeuvring), it is possible to constrain the relative position and separation of each object by modelling the small fractional shift in the centroid of the unresolved point-spread function.
An investigation of this technique has been performed by Dstl and the University of Warwick using data which was obtained from the Warwick Test CMOS test telescope during PHANTOM ECHOES 2. During this experiment, this telescope was used on multiple occasions between February and April 2021 to observe Intelsat 10-02 and MEV-2 over an entire observation night. This included occasions where the two satellites were individually resolvable, where they were not resolvable, and where they were known to have docked together as a single stack.
On nights where each object could be individually resolved within the telescope field-of-view, the collected data could be used to understand the bulk signature and brightness magnitude of each satellite: in general, Intelsat 10-02 would consistently appear significantly brighter than MEV-2; however, during specific glint conditions, MEV-2 would occasionally exhibit a brighter instantaneous response. A representative point spread function was constructed to transform from input point sources to a representational image that was consistent with observations. This model could then be used to arbitrarily alter the relative positions of the two sources and produce a representative output image in each case. By varying the separation of the two sources in the simulated image, particularly to the point where they become merged and non-resolved, and utilising the same detection approach as used with real data, a relationship between the original input positions and the modelled centroid can be determined. From this, a measured centroid can be compared to the model and an angular separation be estimated, such that objects can be tracked at closer (GEO) separation than can traditionally be achieved with the same sensor. This enables improved ability to handle closely spaced objects. A description of this technique is described in this paper, along with examples of application against real data.
During this study, other data collected during the PHANTOM ECHOES 2 experiment was used as ground-truth with which to compare these outputs: this has included contemporaneous data from the Liverpool Telescope, where the two satellites remain resolvable while unresolved using the Warwick Telescope. In this case, the calculated angular separation and direction were found to be consistent. Other sources of ground-truth include data on the targets acquired from other techniques (for examples, passive RF TDOA/FDOA data) which can maintain custody of GEO targets at closer separation.
Date of Conference: September 19-22, 2023
Track: Satellite Characterization