Australia’s Sovereign Conjunction and Threat Warning System

Space debris is increasingly becoming a risk for satellite operators.

Collisions and explosions, coupled with increasing launch rates, create an environment that is becoming more and more hazardous for operational satellites. Possible collisions as well as increased fuel consumption for evasive manoeuvres can put a strain on already expensive missions. 

The aim of Research Program 3 was to create a commercially viable, operational conjunction and threat warning (CATW) system in Australia, with the ability to predict satellite movements and possible collisions in order to warn satellite operators in advance. 

This research was carried out in collaboration with two main partners: Optus, who provided live data on their operational satellites as well as valuable feedback on the generated conjunction data, and EOS Space Systems (EOSSS), who ran tracking campaigns from their sensor sites at Mt. Stromlo, ACT, and Exmouth, WA.

EOSSS also provided basic software code as background intellectual property that was used as the foundation for what became the SERC Conjunction Risk Assessment Program. 

• Conjunction assessment can now be executed on more data sources, most notably SERC’s own catalogue and ephemeris data supplied by Optus. Previously, this capability was available for public Two-Line Element (TLE) data only. 
• Orbital propagators and conjunction assessment code have been ported to graphics processing units (GPUs) to take advantage of their massively parallel architecture. That way, computation times have been brought down significantly. For example, a seven-day conjunction assessment of the entire TLE population can now be executed in under 30 minutes compared to around 8 hours previously. 
• The numerical propagator SENPAI is a complete rewrite and modernisation from a tool that previously existed at EOSSS. While it used to be tightly integrated into the orbit determination process it has been decoupled and made fit for a larger variety of purposes, including conjunction assessment. Like the SGP4 propagator used for TLR data, it has been ported to OpenCL to run on a GPU. 

• Besides being outfitted with GPU capabilities, the conjunction assessment algorithm itself has been modernised. An additional filter was added that substantially lowers the pass rate in the prefilter stage and therefore increases performance.
• With added sigma point generation and the integration with the collision likelihood assessment module (CLAM), another tool created at SERC, conjunction probabilities can now be calculated for objects for which initial covariance information is available. Without parallelisation on the GPU, sigma point propagation would not be possible without a substantial increase in execution time.
• The core algorithms have been put into an application structure that allows for execution as a simple desktop tool as well as a fully-fledged automated service that can be provided to customers.
• Much of the existing software has been modernised, modularised, and/or outsourced into libraries that are easily reusable in other applications.