AI research may shape future defence strategies
Photo L–R: Frank Neumann (University of Adelaide), Tim McKay (DEWC Services) and Claudia Szabo (University of Adelaide)
A team of leading South Australian researchers and defence industry experts from DEWC Services are joining forces to explore how innovative artificial intelligence (AI) and machine learning technologies can power a fleet of simulated self-guided drones to detect and efficiently fight wildfires.
The joint research project, led by the University of Adelaide in collaboration with the University of South Australia, and C4ISREW specialist DEWC Services, was recently awarded $150,000 through the Defence Innovation Partnership Collaborative Research Fund.
As part of the project, a multidisciplinary team comprising of academia, industry and government will apply novel machine learning approaches to a real-world drone firefighting scenario, with plans to then translate the developments into Defence applications.
University of Adelaide Associate Professor Claudia Szabo, who leads the Complex Systems Group at School of Computer and Mathematical Sciences, is leading a team of researchers across multi-agent reinforcement learning, optimisation and heuristics to develop a novel approach that addresses the challenges associated with cooperative AI decision-making in dynamic and uncertain environments.
“Until now, multi-agent reinforcement learning approaches and search-based approaches haven’t been applied in a real-world setting. For multi-agent reinforcement learning approaches, the training time is too long, and search-based approaches are very static and don’t adapt well,” she said.
“The project takes an innovative hybrid approach that combines traditional multi-agent planning and heuristics, and multi-agent reinforcement learning decision-making approaches,” said Szabo.
“The novelty of this approach, is that it is more accurate and adapts better to changes in the environment We know that for Defence, operating in highly complex and contested environments is not uncommon.”
DEWC Services Director – Science & Technology, Tim McKay said some of the company’s brightest engineers will transform the researchers’ novel ideas and technology into a prototype that will be tested and demonstrated in a simulated firefighting scenario using AI-enabled drones.
“We are looking at how these novel AI and machine learning approaches could enable a swarm of drones to work as a team to identify the source of the fire and fight the fire by delivering the exact the amount of water, at the right spot, and do it in the quickest possible time,” he said.
“Working closely with the researchers, we will take these novel ideas and approaches and through coding and prototyping, demonstrate their potential in a scenario. This project will serve as a surrogate for Defence projects. The technology has the potential to greatly enhance the resilience of Australia’s Defence command and control or autonomous operations, particularly in denied or degraded environments. As a trusted partner of DSTG, DEWC Services will seek to translate the technology into existing Defence projects, strengthening Australia’s capabilities in the modern battlespace,” said McKay.
“For example, operating in an environment where their communications have been jammed or they don’t have the ability to share information, so we’re looking at technology that allows swarms of drones to do their tasking somewhat independently.
DEWC Services has a long-standing relationship with DSTG, so we have the touchpoints and ability to apply the developments from this project into classified Defence projects. It enables the collaborative research project to have a higher chance of translating into tangible outcomes that can enhance Australia’s Defence capabilities.”
Related Articles
