From Days to Minutes: How AI Could Transform Anti-Submarine Warfare

Imagine searching for a needle in a haystack, except the haystack is the ocean and the needle is a submarine.

Every day, ships, helicopters, sonobuoys and other sensors collect vast amounts of acoustic data during anti-submarine warfare (ASW) operations. Finding a potential submarine contact within that information can be a complex and often time-consuming process.

For Lieutenant-Commander (LCdr) Mitch Dingwell, who has been overseeing the NEPTUNE project for the past year, the challenge is not simply collecting and reviewing the data but turning it into useful intelligence while it can still make a difference.

"It does not help anyone if we say, 'Congratulations, you found an enemy submarine three weeks ago,'" he said.

"That's great, but we're no further ahead."

That's the problem the Royal Canadian Navy's (RCN) NEPTUNE project is designed to solve.

NEPTUNE is an AI-enabled capability designed to accelerate acoustic analysis and support faster decision-making.

Today, acoustic recordings collected at sea often require extensive manual review. Every minute of sensor data must be analyzed by trained operators, and when multiple sensors are deployed, the workload can quickly grow into hundreds of hours of analysis.

NEPTUNE helps streamline that process by rapidly reviewing acoustic data and identifying patterns that may indicate the presence of vessels or submarines. Rather than requiring analysts to manually examine entire datasets, the system highlights information that warrants closer attention.

Just as importantly, it can help identify what is not a threat.

"If NEPTUNE can confidently identify a contact as a whale or a merchant vessel, the analysts can immediately disregard that data," LCdr Dingwell said.

"You could potentially eliminate hundreds of hours of listening time simply by identifying things that are clearly not threats."

The project has brought together the expertise from across government, industry and academia, including the Centre for Ocean Ventures and Entrepreneurship, Amazon Web Services, Acoustic Data Analysis Centre, and Defence Research and Development Canada. Working collaboratively, the team set out to modernize a process that has remained largely unchanged for decades.

So far, the results have been promising.

During early testing, NEPTUNE generated a report in approximately 15 minutes that would typically require a qualified analyst roughly a day and a half to complete.

What once required hours of manually reviewing recordings can now be transformed into a concise report that identifies potential contacts and provides commanders with a clear assessment of the situation.

Despite its use of AI, LCdr Dingwell stresses that NEPTUNE is not intended to replace operators, but rather to help them focus on the information that matters most.

"Right now, analysts are rotated regularly because after staring at a spectrogram for long periods of time, fatigue sets in," he explains.

"People begin to miss things or think they see things that aren't actually there."

By filtering out irrelevant contacts and highlighting areas that warrant closer examination, NEPTUNE allows analysts to focus their attention where it is needed most, reducing fatigue and improving efficiency while preserving human judgment in operational decision-making.

As the project continues to mature, the vision for NEPTUNE extends beyond shore-based analysis centers.

The long-term goal is to deploy the capability directly aboard RCN ships, providing operators with timely analytical support even when communications with shore are limited or unavailable. In those situations, NEPTUNE could help sailors maintain a high-confidence understanding of the underwater environment while operating independently at sea.

"What I'm really looking forward to is putting this capability onto a ship in a Pelican case and running it alongside existing systems," said LCdr Dingwell.

"I want operators at sea to be able to see the value of it directly."

The project also points toward broader possibilities.

Future applications could include integrating additional operational data sources, while remote sensors and secure cloud infrastructure may one day support near real-time monitoring in remote regions such as Canada's Arctic.

While those possibilities remain on the horizon, LCdr Dingwell emphasizes that the priority remains turning promising concepts into practical capabilities that operators can use today.

"We don't have the luxury of waiting five to seven years for a traditional project timeline," he said.

"We need to put something useful into the hands of operators now and continue iterating over time."

For the RCN, NEPTUNE represents more than a new technology.

It demonstrates how innovation can help transform information into action, reducing analysis timelines from days to minutes, and providing commanders with the timely insights they need to operate effectively in an increasingly complex maritime environment.