Skip to main
/ Log

The dawn of Perseus

Reading time: 12 minutes

POLON-6 at Utah Desert Remote Observatories

If I remember correctly, that time nine years ago I was in the middle of development of few hardware drivers for the robotic observatory located on the top of the roof of Max Planck Institute, Garching, Germany. It was a new, and exciting chapter in my life, the switchover from university times to private sector. Becoming a part of just-launched Polish membership in European Space Agency (ESA) was more than expected and thrilling to me, and I was happy to go all-in. I joined Sybilla Technologies after successfully completing the Founding Project Panoptes PLIIS 1 activity and later on, in Spring and Summer of 2015, took over the development of astronomical algorithms and user interfaces for remote and robotic operations of scientific telescopes. In September 2015, just after the last minute fixes under the umbrella of Moon eclipse, Campus Observatory Garching was officially born. It was the first, commercial installation of Abot (Astronomical Robot), kind of an operating system for the observatory, orchestrating and keeping the hardware secure and conducting the observations, and AstroDrive, a data service designed for keeping and viewing astronomical data.

Laboratory proven prototypes

The following years were packed with ESA projects, research and development in governmental and commercial activities. We focused on Space Surveillance and Tracking domain (SST), that is, tracking satellites orbiting Earth, providing precise measurements and transforming Polish astronomical heritage into the next level: becoming a ground segment for the Space Traffic Management (think of air traffic control, but for space). While ESA projects during that time were crucial for the company survival, they targeted lower TRLs2, up to maybe TRL 4-5, called laboratory proven prototypes, and the operational, practical aspects in the SST domain had to be supported with other sources. In 2018, Poland joined the European Union Space Surveillance and Tracking consortium (EUSST), and our, as a country, goal, was to become a major player in delivering optical data for securing European orbital assets, such as Galileo and Copernicus constellations. From the very beginning, we were supporting Polish Space Agency in that, by building from ground-up National Operating Centre – developing software stack, gathering operational know-how and expertise, training POLSA staff. But the key of success was not only the capability of data processing, building services on top the acquired measurements. The fundamental part was to acquire data at all, and Poland had, and have much to say in that domain.

A telescope inside

If you think of an astronomical observatory, you imagine a dome, a rounded-shape structure, holding inside a telescope on some sort of a tripod (mount), with camera and other scientific equipment attached (camera, filter wheel, focuser). Basically speaking, this is all you need to conduct astronomical observations, and – long story short – satellite observations are not that much different from taking images of nebulas or galaxies. Two things one have to remember here, tough. First, the exposure time is shorter: around a half a second for lower altitude satellites, up to 3-4 seconds for geostationary objects. Second: the satellites are moving faster than speed of sound (much much faster, 8 kilometres per second), crossing the sky within a minute or so. But even that, an astronomy-dedicated observatory like the one we built in Garching, perfectly fits these requirements. Obviously, the faster the mount and the higher the frame rate is, the better, but we’ve shown in one of our ESA projects, Astrometry24.NET, that, for basic SST scenarios, it’s more than enough. And here is why Poland started to shine in EUSST: having dozens of astronomical telescopes around the world (either commercial, like the network of 6Roads, Cilium Engineering, or scientific, like from Adam Mickiewicz University, University of Zielona Góra, Centre for Space Research, Nicolaus Copernicus Astronomical Centre), ready to provide measurements. They were built for different research projects or delivering data for Planetary Defence (tracking asteroids), but eager to be, at least partially, converted into SST ground segment.

Pains and tears

The operational know-how and expertise come in pain and tears however. In sleepless nights, in working after hours, forgetting about sunshine outside, scratching the head each and every day, transfusing coffee into cardiovascular system, glueing everything together with silver tape and prayers. No, really. Even if you pass TRL4, thinking that from now on everything is known, requiring only final touches – it’s quite the opposite. The pain, tears, silver tape and glue stay true. It’s even worse, because you have to show operational, field-tested stuff and there is no place for bullshit.

Gran Canaria from El Teide Observatory

We met our operational pains and tears in managing small network of telescopes, named Panoptes-Solaris, consisting of four Solaris units observing formerly exoplanetary systems, the Open University twins: Pirate and Coast, the Garching veteran, and the new test-bed kid located in Baader Planetarium premises in Mammendorf. The management of one robotic, autonomous telescope is hard, and having more in the bucket is even harder. We contributed with that experience for the ESA’s twins, Test Bed Telescopes, introducing cloud-based sensor network management system, and, later on, extending this platform in the another ESA activity: Web-based tools for SST (aka Webplan). Webplan was the next level, not only in achieving automated scheduling over the whole network for the catalogue maintenance3 scenario, but its heterogenous architecture, allowing us to connect and cooperate with, for example, the laser station from Borowiec. The follow up of the project, P3-SST-XXII aka Polish Small Telescopes for SST proved we can go with that cooperation further, introducing star-and-chase observations, where the data is exchanged between the optical observatory and the laser station for follow-up measurements.

All of these ESA activities reached formally TRL4, in some areas exceeding to TRL6, and, with PLIIS programme going to an end, this was the high time for us to get into the open water.

In 2020 we won a contract for providing two telescopes for POLSA, fully dedicated to SST. Funny thing is that the COVID-19 was everywhere that time, the shortage of electronics was a fact, and we had two months for everything: hardware completion, delivery, deployment and first light. Two sites were selected (one per telescope), in opposite sides of Australia, which, due to pandemic, was closed both for internal and external traffic. We did a two-week installation process and testing in Poland, two-week delivery and a week of remote installation. Fortunately, we ordered fireproof screws required for one site early enough, via an Australian shop, because they were locked for three weeks on a border between South and Western Australia… The first light of each of the telescopes came just before the Christmas holidays deadline and, POLON-1 and POLON-2 were born.

Four days later, on a Christmas Eve, Polish NOC operators got a message of possible collision between Brite-PL scientific satellite and Pegasus rocket debris. The miss-distance, or the distance at the closest approach time, was 4 (four!) meters. Wow, what a luck! In an event of a collision, Polish satellite would be lost. While the POLON-1 and POLON-2 weren’t able to track that, it clearly showed the need for the world-wide network of telescopes focused on tracking traffic in orbit. The idea of upgrading or replacing the ageing astronomical equipment operated by Polish entities raised in POLSA and got the financing from EUSST, but the requirements were pushed to the industry limits.

Have a time to react

Up until now, the status quo of getting measurements for objects in orbit was that the radars take lower altitudes (Low Earth Orbit), and they are efficient in doing so, but at a great hardware and energy cost; lasers can reach lower and some mid altitudes (Medium Earth Orbit), but are confined to a specific set of satellites curated by IRLS, and anything above is a domain of optical observatories (such as Geostationary Earth Orbit, where the telecomm satellites reside). It was usual for our operations to track navigational satellites4, telecommunications and some satellites residing in the upper limits of LEO. The problem in LEO observations though was not not seeing the satellite in the frame, but to have enough background stars to perform automated data processing and object correlation, reaching more than 80% success rate. For this to work, one has to have a really fast mount, really high frame rate camera and a wide field of view, a trio, that wasn’t available off-the-shelf on the market until 2021-22. At the same time, we observed a dawn of satellite constellations, such as Starlink, adding more than 50 new objects in orbit per week. We observed collisions in space or near misses. We observed intentional destruction of satellites with a rocket launched from a military base. All of them participating in increased number of close approach events, increased number of space debris occupying LEO, endangering Earth observation, human presence on space stations, and – since we are all dependent now on satellite-based services – affecting our daily lives. With that in mind, space traffic management no longer is a science-fiction and requires ground-based hardware of different kind looking at the night sky continuously, processing and delivering data near real-time (current limit for EUSST applications is 48 hours), so that decision makers, satellite operators will have time to react.


The cost of a whole network of high-end optical telescopes is lower than a building a single radar, and it was decided by the Polish Space Agency to invest in such, during the round of hardware upgrades in the EUSST consortium. The tenders for extension to POLON-1 and POLON-2 by 5 new sites were published, and we won 4 of them, proposing a setup not of 4, but 16 telescopes in total, with the fifth site developed by Cilium Engineering following that idea. The timeline was harsh and strict, observational metrics hard to believe that are achievable (tracking really small objects in LEO5), but we closed 2023 with 12 new telescopes ready to work, with 8 more, at the time of writing, nearly completed.

Once finished, the network will consists of 20 telescopes in 5 locations: POLON-3 in Siding Spring Observatory (Australia) POLON-4 in Deep Sky Chile, POLON-5 in South African Astronomical Observatory, POLON-6 in Utah Remote Desert Observatory and POLON-7 in Hawaii. Each site contains 4 similar units based on PlaneWave L-350 mount, ASA UWF300 optical tube and QHY411M Pro camera, orchestrated by Abot, with on-board data processing via Astrometry24.NET and Tracking Data Messages6 delivered within 2-3 hours. The telescopes can work separately, tracking different objects in all orbital regimes, or may work together, surveying a vast part of the night sky at the same time. This setup is great not only from the observational perspective, but also from the redundancy point of view, similar to Space Shuttle computers or engines. Even if three of four units fail, there is still one left providing data to National Operating Centre, and EUSST for services such Collision Avoidance7. This uninterrupted software and hardware service is not just a contract requirement, but sets a high bar, a new standard for operating a network of telescopes, measurement precision, data processing and delivery, and shifts us from laboratory-proven prototypes (TRL4) to flight hardware (TRL9). We called it Perseus, after a mythical hero, slayer of the dreaded and deadly Medusa. It is a new and exciting time for the company, redefining multiple building blocks, gathering know-how into a streamlined products, staging from the research and development projects we did over the past decade.

Keeping an eye

I was overseeing the deployment of 4 sites of the POLON network, being personally involved in the installation of 3 of them, travelling around the globe, co-managing the project, testing hardware and software, finally, keeping an eye on everyday operations. It was hard, stressful, exhausting but at the end of the day – a great achievement of the whole team! I hope that the launch of Perseus will strengthen our activities in the Space Safety domain, because we’ve already shown what optical systems are capable of.

One particular example of that is the picture above. If someone told me few years back that we will be able to track 12x5x5cm space debris at 600 km orbit (it’s like 5 iPhones glued together, or a half of a kilo of sugar, travelling at 8 km/s and having the impact energy of a fully loaded truck), I wouldn’t believe in that, unless pictured. More than that, with the fantastic performance of the Robotic Telescope of Zielona Góra (ROTUZ aka Panoptes-4) and Agnieszka Being Observer Tonight, we reached lower limits for the optical observations, tracking the final stages of ESA Aelous mission, hours before its reentry. Our final image was taken when the satellite was at around 260 km, and we would picture it at 160 km if the observatory roof allowed us to see 4 degrees above horizon.

And there is more to come. While the beginning of Spring 2024 brings us the off-the-shelf Perseus system, work is also focused on developing European Optical Network, a TRL7 demonstration of the end-to-end process for keeping an eye precisely on satellites. Stay tuned!


  1. PLIIS: Polish Industry Incentive Scheme, a 7-year ESA programme dedicated to developing the Polish space industry readiness for working and competing in the European market. 

  2. TRL: Technology Readiness Level 

  3. Catalogue Maintenance: keeping a set of satellites in safe orbits 

  4. Global Navigational Satellite System 

  5. Cubescout-S1, 12x5x5cm at 600km orbit 

  6. Tracking Data Message is a standard in the Space Safety domain for exchanging measurements of satellite postion and brightness 

  7. Collision Avoidance: a satellite manouver required for avoiding crash with another satellite or space debris