Artificial Intelligence and Maritime Maneuvering - Collaborative Design and Competition with NSWC Crane
Abstract
Trine University will be undertaking two overarching projects through the Program Increase - Artificial Intelligence Maritime Maneuvering (AIMM) initiative.The first project will be to develop a Low Profile Vessel (LPV) design in partnership with NSWC Crane that can be easily produced with limited fabrication skills and equipment; as well as, easily obtainable materials (e.g. wood, PVC pipe, plastic, and fiberglass) from general consumer supply chains. The LPV design will be developed as a small unmanned surface vessel (USV) of approximately 8-12 feet in length with the ability to easily be modified to deploy a variety of marsupial devices (e.g. an Unmanned Aerial Vehicle #UAV#, Unmanned Underwater Vehicles #UUV# or other Unmanned Monitoring Systems). The skeletal design will beeasily accepting of a variety of drive systems (e.g. gasoline, diesel, or electric) and will have the capability to house a yet to be determined sensor package and microprocessor system. As a portion of thedesign refinement, Trine University will collaborate with NSWC Crane to field test a variety of sensors to best determine what sensors will work for on sea data collection. Trine University will be putting the design concept through theoretical computer simulations utilizing finite element analysis (FEA) and computational fluid dynamics (CFD) modeling; as well as, physically testing scale models in lab experiments. The preliminary full-scale prototype will be field tested in regional lakes surrounding Trine University#s campus in Northeast Indiana. The culmination of this design collaboration will be a simple LPV technical drawing and bill of materials kit to be leveraged as a baseline design for five Indiana-based colleges and universities to compete in an Unmanned Artificial Intelligence Surface Vessel competition to be hosted in May 2024. For the second parallel path project, Trine University will develop an Artificial Intelligence Multi-Agent System at Sea (MASS) Swarm Intelligence Training Simulation (SITS). This project aims to answer the question: how can a team or #swarm# of small LPVs best cooperate, track, and capture defined combatant vessels? The two central themes of this project are autonomous cooperation and swarm intelligence. In addition, Trine will aim to create a variety of intelligent agents that are capable of accurately simulating a variety of ships at sea per inputs provided by NSWC Crane. This visual AI Training Sim will include but not be limited to ocean traffic patterns, ocean currents, commercial vessel specifications, private vessel specifications, DOD vessels, etc. Trine University will directly collaborate with the Office of Naval Research to include knowledge and capabilities already developed within the Navy#s #AI Captain#. The sim will be tailored to contain known inputs pertaining to previously apprehended LPVs and southern coastal waters specifically pertaining to anti-narcotic trafficking. The final training sim#s goalwill be to create an advanced AI Multi-agent system of autonomous boats that are capable of safely and efficiently achieving the Navy#s mission in this environment. The sim will provide an expedited learning environment to train future artificial intelligence platforms to be deployed for counter narcotics efforts.All of the above referenced project tasks and objectives will be completed by a project team of university faculty andstudent interns from Trine#s Allen School of Engineering and Computing. These faculty and students will be overseen by the University#s external partnering department, Innovation One.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jul 24, 2023
- Source ID
- N000142312661
Entities
People
- Jason Blume
Organizations
- Office of Naval Research
- Trine University
- United States Navy