The US Navy has awarded Charles River Analytics $1 million to develop a tool that uses artificial intelligence (AI) to identify adversary submarines by their magnetic signals.
Unlike traditional acoustic detection methods, the Guidance for Naval Enhancement of Tactical Operations (MAGNETO) tool capitalizes on the fact that submarines, made of ferromagnetic materials, distort the Earth’s magnetic field.
However, detecting magnetic fields alone is insufficient for effective anti-submarine warfare, the firm explained.
It’s equally important to distinguish submarine signals from those of large boats or sunken shipping containers. AI plays a critical role here by enabling the creation of unique “signatures” for submarine detection.
“The biggest advantage of MAGNETO is that it allows us to detect ultra-low-noise submarines reliably, quickly, and efficiently while minimizing risk to the warfighter,” said Dr. Todd Jennings, Research Scientist at Charles River and Principal Investigator of MAGNETO.
How MAGNETO Works
MAGNETO employs a hierarchical process, progressively refining signal identification through successive stages.
This structured approach allows the system to process data in real-time, ensuring that only relevant signals advance to more detailed analysis.
Initially, MAGNETO detects the presence of a submarine, followed by determining its type — such as a nuclear missile or diesel-powered — until achieving a precise classification.
Incoming signals are first cleaned to remove noise, after which a feature extractor isolates key elements that might indicate activity.
In the final stage, a classifier conducts a detailed analysis to confirm the submarine’s identification.
Future iterations of the tool are expected to further enhance its capabilities by analyzing wave noise to determine if it signals the presence of submarines.
Beyond submarine detection, the tool can also locate airplane wreckage, underwater power cables, and lost shipping containers.
