DURIP: Acoustic Sensing System and High-Throughput Computing for Environment and Threat Monitoring in Naval Environments Using Machine Learning
Abstract
"This proposal is a request for equipment to primarily support two active U.S. Navy projects: ONR Graduate Traineeship Award: Multi-Modal, Near-Shore, Ice-Covered ArcticAcoustic Propagation Measurements and Analysis (ONR #N00014-18-1-2592) andLocalization, Tracking, and Classification of On-Ice and Underwater Noise Sources UsingMachine Learning (US NSWC #N00174-19-1-0004). This equipment will also significantlyenhance our capabilities for six other Department of Defense (DoD)-funded projectsNGA, SPAWAR, and DARPAand will augment Navy STEM education and workforce development through the SENSE (Strategic Education through Naval Systems Experiences) Enterprise at Michigan Tech University. Specifically, we are requesting i) new high throughput computing (HTC) equipment and ii) two headless underwater acoustic sensing systems consisting of a data acquisition system, hydrophones, and vector sensors. Our request combines headless acoustic data acquisition systems intended for sparse sensing combined with HTC facilities to implement machine learning and AI approaches for target tracking in complex environments, particularly ice-covered seas. The HTC and sensing system equipment will be used to accelerate research in detection and localization of on-ice and underwater anthropogenic noise sources with deep learning; sensing, prediction, and path-planning in autonomous underwater vehicles and surface vessels, and detection and classification of targets in overhead imagery using deep networks. This equipment will also enhance ongoing collaborations with industry in the areas of signature modeling and observability prediction, acoustic sensing and processing, acoustics measurements and characterization, and adaptive array processing. Furthermore, the requested equipment will enable expansion of current research into numerous related areas of interest to the DoD, such as deep learning with digital phased arrays and persistent, distributed sensing with sensor arrays.We have excellent resources, but lack infrastructure capabilities outlined in this proposal to make a leap in theory and applications and conduct more detailed, realistic, and repeatable sensor/target experiments on past, current, and future U.S. Navy data, and Department of Defense (DoD) data, overall. Furthermore, this DURIP proposal connects experienced investigators from different research projects. As a result, the Navy will get a value-add by knowledge and code exchange. Techniques developed using this equipment in current and future projects can be extended to many DoD-relevant domains."
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 31, 2020
- Source ID
- N000142012793
Entities
People
- Timothy C Havens
Organizations
- Michigan Technological University
- Office of Naval Research
- United States Navy