DAHLGREN, Va. –
Since it’s fielding in World War II, radar has produced crucial early warnings of incoming enemy forces and better accuracy against targets, giving the U.S. Navy a decisive advantage in naval warfare engagements.
Nearly eight decades later, radar – which uses electromagnetic waves to detect and track objects – remains a cornerstone of modern defense systems. What began as a basic detection system has evolved into sophisticated system of systems that can detect, track and engage stealthy, fast-moving targets and provide increased lethality helping safeguard critical assets.
Still, a fundamental challenge endures — one as old as radar itself. Although radar can detect the presence of an object, it cannot inherently identify it, making it difficult to separate genuine threats from “clutter.” Clutter refers to unwanted reflections that can obscure the presence of an object from the radar system. These unwanted reflections can be due to moving objects such as wave caps, birds or precipitation, or can be due to complex reflections from static structures such as buildings (point clutter) or a combination of all.
Research underway at Naval Surface Warfare Center Dahlgren Division is working to find a new, foundational way to look at – and hopefully resolve – the issue of radar clutter. The work is part of the Naval Engineering Education Consortium, a Naval Sea Systems Command Warfare Center-directed initiative that involves college and university students in project-based research to address the Navy’s most pressing technology needs.
“This research in particular aims to develop automated, accurate and robust machine learning algorithms to mitigate radar clutter in shipboard systems,” said NSWCDD NEEC Director Caleb Strepka. “This will ultimately enhance the U.S. Navy’s situational awareness and transform large datasets into a strategic advantage by significantly improving target detection, tracking and engagement in the presence of clutter.”
NSWCDD has collaborated with Dr. Justin Krometis, a research associate professor in the Intelligent Systems Division of the Virginia Tech National Security Institute, and Will DeStaffan, a computer science graduate student at Virginia Tech.
“In a naval defense scenario, clutter is the biggest challenge faced by radar operators,” DeStaffan said.
The goal: Make on-ship radar clutter mitigation as efficient, accurate and fast as possible.
“We also want to make it explainable. When you have a radar operator sitting on a ship in an air defense scenario, they want to know and trust the system is making the right decision – and know why it made the decision it did,” DeStaffan said.
The stakes are high. Misidentifying a cloud as a threat could trigger unnecessary defensive actions. Likewise, failing to detect a real threat could leave a ship vulnerable to attack. The situation becomes even more complicated in complex environments like at sea where the ocean can cause reflections that can distort a radar signal.
The research looks at whether a solution lies in artificial intelligence and deep learning techniques by using months and even years of data.
“This is a very new and exciting field to apply to this decades-old problem,” DeStaffan said.
Before he began his research in the summer of 2023, DeStaffan never imagined getting into radar work. He was more interested in drones and complex communications systems. When a promised internship fell through, a professor put him in touch with the head of Dahlgren Division’s Advanced Radar Systems Branch in the Electromagnetic and Sensor Systems Department.
The more he worked on the project, the more interested he became in finding a solution.
The work matters, DeStaffan said. “Once we get this to a place we want to get it, and it’s shipped off to a boat or to Dahlgren for analysis, this is an additional tool that’s supporting the warfighter, which in turn is supporting the country.”