Safe Maritime Autonomous Systems

Offshore wind farms, subsea pipelines and coastal defence systems require constant monitoring and maintenance. Historically, this has meant deploying divers and remotely operated vehicles, methods that are costly, labour-intensive and carry significant safety risks. Maritime autonomous systems (MAS) offer a powerful alternative to make offshore inspection faster, cheaper and safer. This project is exploring how to realise this potential responsibly by deploying underwater sensors and autonomous vehicles as an acoustic mesh network, and developing lightweight machine learning models suited to autonomous maritime operations alongside the rigorous safety assurance frameworks needed to deploy them in high-stakes environments.

What sets this project apart is its interdisciplinary approach, with the research team uniquely drawing on expertise from across all three of ISA’s core research pillars:

• Communications: Leveraging expertise in underwater communication, sensing systems, and terrestrial/high-altitude 5G/6G networks to keep autonomous systems connected.
• Design & Verification: Building the physical and digital architecture for autonomous underwater vehicles (AUVs) and marine robots.
• Assurance: Applying frameworks like AMLAS (Assurance of Machine Learning in Autonomous Systems) and SACE (Safety Assurance of Complex Environments) to map hazards, define mitigations and ensure reliable decision-making under severe communication disturbances.

A major highlight of the project has been a public workshop bringing together over 50 experts from industry, academia, and government to discuss the future of safe maritime autonomous systems.

The event featured a live demonstration in ISA's anechoic tank of NEMO, a prototype device that connects underwater sensor nodes, AUVs and surface buoys into a wireless acoustic mesh network. The aim of the NEMO system is to enable real-time monitoring, control and positioning of subsea assets. This includes the development of a novel acoustic communication protocol stack that allows the telemetry, control and positioning signals to coexist and operate concurrently on the same acoustic network.

Read more about the workshop

Project leads: Nils Morozs, Tian Gan

Other examples of leading research carried out by the Underwater Information Systems group include:

• Advanced signal processing for high data rate, multi-user acoustic communication
• Combined acoustic communication and underwater target detection
• “Full-duplex” (simultaneous transmit-receive) acoustic communication and transducer design
• Underwater acoustic network protocols, including node discovery, medium access control and routing algorithms