Professor Tat-Jun (TJ) Chin
School of Computer and Mathematical Sciences, University of Adelaide
SmartSat CRC Professorial Chair
Defence Trailblazer Theme Lead for Defence Space Technologies, University of Adelaide
Professor Tat-Jun (TJ) Chin has forged his career around developing state-of-the-art solutions for Space using innovative Artificial Intelligence (AI) techniques.
As the co-Theme Lead for Defence Space Technologies, he contributes his foundational research expertise to uplift the Australian defence space industry and achieve commercial impact.
From AI fundamentals to industry leader
“Early in my career, my focus within AI was mainly on the fundamental level. My foray into Defence came after engaging with companies that work in the defence sector,” said Professor Chin.
Since the establishment of Defence Trailblazer, he has been instrumental in constructing projects with multiple industry partners. “What I find most rewarding about this role is being in a position to influence the running of a major program like the Defence Trailblazer,” he said.
“Through Defence Trailblazer, I wish to contribute to changing the perception of industry-led research in the academic sector. Many academics do not realise that industry-led research can be very rewarding, and achieve a path to impact beyond publications, grants and citations.”
AI and Space capabilities
Professor Chin’s research interests include the optimisation of computer vision and machine learning, with applications for intelligent satellites and space robotics.
“Within the massive area of artificial intelligence, my interest is mainly in perception algorithms – algorithms that make sense of visual data. Visual data includes images, 3D video scans, and other data generated from more sophisticated sensors. These have a lot of relevance in Defence applications.
The research that we do at the University of Adelaide is centred around the concept of what I call ‘AI for Space’. There are two major strands of activities, or subthemes – one is autonomous processing of large-scale data, and the other is autonomy for Space systems.”
Autonomous data processing
“Defence systems need to make sense of the surrounding environment for decision-making, across land, sea and Space. Often some kind of sensor is used that collect data from the environment, such as electro-optic sensors for situational awareness and remote monitoring. Even if the sensors don’t directly produce visual outputs, the data can be converted into a visual form.
In the case of Space, satellites fly around the Earth, observing and taking measurements of the Earth’s surface. Then there are sensors that point towards Space to detect, track and surveil objects of interest. All of this produces visual data that needs to be processed.
Manual processing of all the Space, Defence and civilian data is impossible, so we need algorithms to process and extract the value of the data.”
Autonomy for Space systems
“The subtheme of autonomy in Space systems is a newer area that I am helping to pioneer globally. This research area is becoming more important as space applications become more and more complex, with a need to perform more and more sophisticated tasks, such as repairing satellites in Space.
Often these tasks take place far away from Earth, such that manual remote control is not possible. Alternatively, they need to occur in split second, or they need to occur over long periods of time that it becomes infeasible for humans to remotely control and babysit all these. Therefore, we need to give these systems some level of ability to make decisions autonomously and automatically.”
Research leadership
“I lead an AI for Space Group of research associates and students, with about half of the research associates working on projects related to Defence.
As a Theme Lead for Defence Trailblazer, I provide academic input for the programs, in terms of the direction of R&D to take and also the nature of the partnership between universities and industry.
My other responsibility invokes my existing relationships with industry partners who work in the Defence sector – bringing them into the Defence Trailblazer fold so they can engage with us concretely.”
R&D for impact
“Defence Trailblazer provides a great platform for industry partners to access university R&D capabilities. That’s the baseline benefit.
The add-on benefit is being able to match industry offerings with potential customers, usually in Defence. The benefit to industry partners is that both industry and Defence Trailblazer provide matching funding to share the cost of R&D – as well as gaining that important connection to end users.
As for the benefits for academics, these programs suit academics who want to generate the kind of impact that leads to commercial outcomes. After all, the end goal of the Defence Trailblazer program is to solve problems that really matter to industry partners and their end users.”
Background
Professor Chin is the University of Adelaide Professorial Chair of Sentient Satellites. He received his PhD in Computer Systems Engineering from Monash University in 2007, which was partly supported by the Endeavour Australia-Asia Award.
He has published more than 150 research articles on the subject, and nd has won several awards for his research, including a Computer Vision and Pattern Recognition (CVPR) award in 2015, a British Machine Vision Conference (BMVC) award and Best of European Conference on Computer Vision (ECCV) award in 2018, two awards from the Defence Science and Technology Group in 2015 and 2017 respectively, and an International Association of Pattern Recognition (IAPR) award in 2019.
Professor Chin is also the Director of Machine Learning for Space at The Australian Institute for Machine Learning. He was a Finalist in the Academic of the Year Category at Australian Space Awards 2021.
