Currently, SERDP and ESTCP are supporting efforts to cost-effectively recover and remove munitions in the underwater environment. Such practices usually involve divers to manually retrieve targets, making it a dangerous and expensive endeavor. Additionally, most underwater sites identified as potentially containing munitions are in shallow water (0-120 feet), which poses a greater risk to the public. For this reason, it is important to focus on safe and efficient methods to remediate underwater munitions that cannot be moved for explosive safety concerns.

SERDP Exploratory Development (SEED) projects test a proof of concept over the course of approximately one year. In 2021, SERDP funded five SEED projects aimed at advancing technologies that detect, classify, and remediate military munitions at a variety of underwater sites. A particular need in this area of research is to improve wide area and detailed surveys that would more quickly characterize, map, and detect munitions under complex conditions. Projects also address sensor development, platform integration, analysis methodologies, and large-scale collection of field data at real munitions sites.

• At Penn State, Dr. Vishal Monga is leading a team to develop a deep learning approach using data from an existing sonar system developed under a previous SERDP project that produces 3D synthetic aperture sonar (SAS) imagery to detect and classify unexploded ordnance (UXO). Deep learning frameworks have great potential for detecting UXO, but they significantly rely upon available training imagery. This project will help address this challenge and build a foundation for deploying machine learning to remediate UXO (Project Overview).
• Dr. Meagan Wengrove at Oregon State University and her project team will use small-scale physical modeling to classify the vulnerability of munition migration, exposure, and burial on beaches within the surf-swash transition zone during dam-break forcing conditions. While previous SERDP projects have performed experiments on sandy beaches and muddy bottoms, this effort will specifically focus on mixed-grain (sand and gravel) environments, which make up 24% of Formerly Used Defense Sites (Project Overview).

• Dr. Suren Jayasuriya at Arizona State University and his project team will provide new algorithms for improving beamforming and automated target recognition (ATR) of UXO in cluttered and buried underwater environments. They will merge physics-based knowledge of munitions (scattering models, target modeling, and environmental characterization) with machine learning-based ATR algorithms (Project Overview). 

• Autonomous underwater vehicles (AUVs) with high-resolution seafloor imaging sonar could provide a more efficient and safer means of surveying UXO sites, but such tools depend on robust autonomous detection and classification algorithms. Dr. Alan Hunter at the University of Bath will develop a data simulation capability that provides more labelled training data by increasing the speed of SAS data modelling. The team’s goal is to prove the concept and use it as the foundation for a follow-on proposal to develop a fully capable 3D simulator (Project Overview). 

• At Woods Hole Oceanographic Institution, Dr. Peter Traykovski and his team will focus on improving hydrodynamic forcing models of UXO migration and burial in shallow water surf-zone environments. Numerical modelling of environmental conditions can help predict UXO mobility and burial throughout shallow water regions. This project will improve UXO management by providing more knowledge of UXO migration during wide area surveys, detailed surveys, and removal (Project Overview).

These SEED projects allow for greater exploration into novel and emerging technologies that have the potential to advance UXO remediation efforts. To learn more about SERDP and ESTCP munitions response research, browse other projects here.