Engineering Insights: A Conversation with Martin Cupak
Dobré ráno
It’s a bright morning as I traipse across the campus. Clouds scuttle across the sky. I scurry into building 314, my feet pounding the carpet and beating a rhythmless tune. This time, I know where I am going, both the room number, 166, and the title, Desert Fireball Network headquarters. The door opens and I am greeted by my interview subject for today—Martin Cupak from the Czech Republic.
The Computer Says Martin
‘My background is computer science. I have a master’s degree in computer science. I’ve been doing a big variety of software work from embedded systems, assembly language, machine language, to higher level languages and working with big data supercomputers and a bit of everything in between.’ If there was one person at DFN who represents a modern, tech-savvy version of MacGyver, Martin could fill the role. Since 1999, he’s been working with ‘automated instruments for photographing the night sky to capture fireballs that might drop meteorites.’ Martin started out in the Czech Republic, and when their instruments proved to be a success, they deployed the cameras in Australia. He has been part of the DFN team with Curtin’s Space Science and Technology Centre (SSTC) since 2013. As a valued team member, he is across ‘a lot of different engineering things to support the project, from coding, like writing software, to field work, fixing instruments in the lab and in the field.’
Teamwork makes the Dream Work
Martin is a key software engineer for the DFN. His coding and programming ensure that the team can analyse relevant and accurate data in a timely manner. However, he is not the only programmer, and he made sure I understood this. He took me through a detailed walkthrough of what goes on behind the scenes at the DFN, detailing the experience of when a fireball enters the atmosphere to how the scientists extract data on possible landing spots.
‘The fireball can happen at any time. In the morning, the images are processed with an algorithm that detects what might be a fireball. That runs on each of the camera systems, which has its own small PC computer. The results of this detection are transferred to the server, which is underneath a desk at the DFN office. The server pulls high-resolution images from the cameras that detect something and collects that into a folder. Then the processing happens, just producing files in that folder for the fireball event.’
At Base
‘The fireball is seen from different cameras’ point of view, so you get a sort of curve in 3D, point by point. Once you have that, then you can run other algorithms like analytics of the flight, how it was decelerating, what the angles were.’ There are other calculations that they can run including what was the initial mass of the object upon atmospheric entry to the end mass. ‘That streak in the sky that the fireball makes, that ends some 30, 20, 40 kilometers high.’ At that point, the fireball is no longer visible to DFN cameras, but it keeps falling to Earth, known as its dark flight. ‘If we want to find a rock, we need to simulate the dark flight.’ All the simulation software is designed and coded by the team Martin works closely with. ‘My part is gluing all this together, organising the data transfers into those steps of processing. I’ve been involved in the weather modeling and automating that and a little bit in the embedded software and the cameras.’ Martin is also involved in ‘the maintenance of automatic updates if we make some change in the software, so it propagates to all the cameras, and keeping an eye on the cameras that actually operate, or if there’s something wrong, trying to analyse what’s wrong.’
If we want to find a rock, we need to simulate the dark flight.
In the Field
Martin has an exciting job as an engineer in the field. ‘One part of field work is servicing the instruments, the camera systems, and then the other type of field work is meteorite searching.’ Both roles give him the opportunity to ditch the office and hit the bush. This is why Martin prefers to be known as an engineer, rather than a programmer. It’s about the mindset, not the milestones or the mileage. ‘The concept of engineering is to break down things into smaller subsystems. Then you can focus on the one thing, and you fix that, and then you go up again and check whether it works.’
‘The DFN makes it possible for me to do quite a lot of fieldwork. We have deployed 50 camera sites across Western Australia and South Australia. I was pretty much on every one of those sites, and on most more than once.’ Martin enjoys fieldwork because it gives him an opportunity to visits places where one wouldn’t normally choose to go for holidays. And as he keeps going there, he appreciates the specific character of the remote country. While in the office, I saw some pictures where cameras were positioned. There are no roads, just tracks, if you could call them that. The work is not glamourous, but it is fulfilling. ‘Usually, you work 12 hours a day in the field.’
Cameras and Communication
Each rugged little DFN station is a self-sufficient science outpost in the middle of the Australian outback. They’re solar-powered, with panels mounted on a sturdy metal frame out in the open desert, soaking up the sun. ‘The panels provide power the equipment and charge batteries during the day, then the battery keeps the camera running through the night.’
The cameras must communicate with the DFN server, back in Perth. ‘Most of them have a mobile data network, like a Telstra Connection.’ The connection can be through either mobile or satellite networks, but use a low bandwidth. It would be too expensive to transfer all the data captured by the remote cameras. Once a fireball is spotted, only the specific subset of data is automatically uploaded so the DFN team can perform triangulation and analysis across multiple stations.
Decommission
Some of the cameras are due to be decommissioned in the future. This is due to the high maintenance and personnel costs associated with running them. As a university research group, the DFN runs on funding from various sponsors and grants, but sometimes the life of a project can extend far beyond what was expected – and it can be difficult to get new funding for an long-running project. ‘When it comes to science, you need technology and equipment. Without funding, you can’t get that.’
Martin must get back to work. I thank him for his time answering my questions that took me on a deep dive into the camera systems, a science I had overlooked in previous interviews. Without the cameras, the DFN is blind. And without the software to analyse the vast amounts of data the cameras collect every night, there wouldn’t be enough people to work through each photo every day.
Written by Louise Kaestner, 2025. For more, visit LinkedIn or head to her website: https://www.louisekaestnerwriter.com/