This project explores autonomous swarms for detection of underwater objects using decentralised decision support systems and signal detection using machine learning. Sensing, decision, and control of maritime systems are a complex set of tasks due to challenging ocean environments. There is a growing trend towards supporting or replacing high-value crewed vehicles with low-cost autonomous/un-crewed vehicles which can be deployed in greater numbers. In the maritime environment, one particular advantage of using swarms of un-crewed vehicles is to achieve multiple sensing pathways, which can improve sensing performance (through both redundancy and sensor fusion) and allow more flexible decision support systems. However, there are also challenges associated with the communications and control systems required for such multi-agent systems.
The project aims to investigate multi-agent strategies for swarms of maritime vehicles, to investigate opportunities defined by: distributed array sonar-like sensing algorithms for sensing underwater objects and reducing noise exposure to the environment; optimal control strategies for locating units within the maritime environment to improve sensing performance; as well as, multi agent sensor fusion for enhanced autonomous marine navigation, situation awareness, decision making, communication and coordination.
The project is also expected to produce a generalised multi-agent simulation tool, which will include customisable physics components allowing sensing, communications, and control systems models to be tested via monte carlo simulation approaches.
Industry Partner: Saab Australia
University: The University of Adelaide
Academic Supervisor: Dr. Will Robertson
Email: [Email Address Removed]
Offered for: Doctor of Philosophy (PhD)