ARC research grant to safeguard UAV technologies

Charles Sturt University’s Artificial Intelligence and Cyber Futures Institute (AICFI) has received funding from the Australian Research Council (ARC) for a pioneering research project titled ‘Robust defences against adversarial machine learning for UAV systems’. The ARC grant is $445,000, with approximately $100,000 apportioned to the university.

The initiative is a collaborative effort between the AICFI, the University of Wollongong and the University of Southern Queensland, and aims to explore resilient cybersecurity measures for unmanned aerial vehicle (UAV) systems to safeguard them against adversarial machine learning attacks. The AICFI lead for the project is Dr Fendy Santoso, who will be responsible for implementing and evaluating countermeasures to implement the defence mechanisms developed and to integrate them with real-world UAV systems using datasets provided by the partner organisations.

“Once deployed on the UAV systems, the defence countermeasures will be evaluated in terms of their effectiveness and real-time performance,” Santoso said.

“We intend to assess a wide range of UAVs. For example, fixed-wing UAVs, which are bigger and more sophisticated, will be deployed in remote areas and highways. In comparison, rotorcrafts and quadcopters, which are smaller with limited speeds, can be used in urban areas.”

Santoso explained that UAVs play a pivotal role across national industry sectors such as logistics, environmental monitoring, smart farming, and bushfire and disaster management. “However, the susceptibility of machine learning models on UAV systems to adversarial attacks poses a significant barrier to their widespread adoption,” he said.

“The project seeks to advance our understanding of cybersecurity by employing innovative approaches to fortify UAV systems against vulnerabilities in machine learning models.”

The project’s anticipated outcomes encompass refined techniques for developing robust machine learning models and an increased capability to design secure UAV systems. These advancements are expected to significantly enhance the security of UAV technology and foster reliable use in transport and logistics services to support Australian urban and regional communities.

“We look forward to advancing our collective knowledge in cybersecurity and contributing to the secure evolution of critical technologies,” Santoso said. The project will commence in 2024.

Image credit: iStock.com/evandrorigon