MichaelBenjamin
Principal Research Scientist, MIT
Triton Summer School
Marine Autonomy
Based on MIT course 2.680 — Marine Autonomy, Sensing and Communications
Design, develop, and field-deploy the autonomy that runs on a marine robot. June 15–26, 2026.
The opportunity
MOOS-IvP is the autonomy middleware behind MIT's marine robotics research and a growing number of defence and commercial USV and UUV programmes worldwide. Engineers who can configure, deploy, and extend it are rare — and demand is accelerating as autonomous maritime systems move from prototype to operational. MIT 2.680 is the fastest path to that proficiency: two weeks moving from simulation to competitive multi-vehicle field operations with real hardware. The skills transfer directly to UUV programmes, larger autonomous platforms, and any context where you need to specify or evaluate marine autonomy systems.
Why this course, why now
The procurement and research timelines for autonomous maritime systems have compressed. HCDI 2026 calls, EDF programmes, and NATO maritime autonomy requirements are all creating near-term demand for teams that can write technical specifications, run simulations, and operate autonomous marine vehicles in the field. MIT 2.680 covers all three. Two weeks is enough to develop genuine operational proficiency — simulation competence, field deployment experience, and MOOS-IvP fluency — rather than a certificate that documents attendance. The organisations that send engineers now will have documented, credentialled capability on record before most regional competitors have begun to address the gap.
What you will learn
Survey of the principal approaches to autonomous behaviour generation: reactive, deliberative, and hybrid architectures. How MOOS-IvP implements a decision-making layer for marine vehicles in practice.
Hands-on development with the MOOS-IvP open-source platform: process communication, behaviour authoring, mission configuration, and debugging in simulation and on hardware.
Protocols and architectures for coordinating multiple autonomous marine vehicles: acoustic communication constraints, task allocation, deconfliction, and formation strategies.
End-to-end mission design: objective definition, vehicle tasking, runtime monitoring, post-mission log analysis, and how to improve performance across iterative deployments.
Practical skills for in-water operations with autonomous surface vehicles: payload autonomy computer assembly from basic components, pre-deployment checks, safety protocols, abort procedures, real-time monitoring, and post-mission recovery.
Design and execution of multi-vehicle missions in competitive scenarios — the core of Week 2. 1v1 and 2v2 competition formats modelled on MIT's own course competitions. Strategy, coordination, and real-time adaptation.
How the course works
Week 1 is simulation-intensive. Week 2 is hardware and water.
Week 1 — Simulation and Architecture
Architecture overview. Simulation environment setup. First MOOS processes. IvP behaviour authoring. Single-vehicle mission exercises in simulation.
Multi-vehicle communication models. Task allocation approaches. Deconfliction strategies. Simulated two-vehicle coordination missions.
Mission log analysis. Performance metrics. Iterative improvement from simulation results. Introduction to hardware interface.
Full simulation missions. Transition briefing for Week 2 hardware operations. Safety protocols for in-water deployment.
Week 2 — Hardware and In-Water Operations
Porting MOOS-IvP behaviours from simulation to real BlueBoat USVs. Initial in-water single-vehicle missions. Hardware debugging.
First in-water multi-vehicle coordination missions. Real-world acoustic communication testing. Iterative tuning.
Three-day competitive finale: 1v1 autonomous missions and 2v2 team coordination. Teams design strategy, implement behaviours, and compete. Faculty judging. MIT Open Learning certificate ceremony.
Course Highlights
Who should attend
The programme is designed for engineers and scientists ready to move from conceptual understanding to hands-on operational competence. Typical participants include:
- Marine and naval engineers seeking direct experience with autonomous surface and underwater vehicle systems.
- Robotics and systems engineers expanding into maritime applications (from aerospace, ground robotics, or industrial automation backgrounds).
- Defence R&D professionals preparing for procurement, proposal, or program management work involving unmanned systems.
- Systems integrators evaluating marine autonomy platforms and software frameworks.
- Researchers in ocean science, environmental monitoring, or offshore energy seeking to add autonomous vehicle capabilities.
- Technical team leads building organisational competence in marine robotics.
- Engineers and scientists from organisations planning to respond to HCDI R&D 2026 calls or similar European defence and maritime innovation programmes.
Expected technical background
Aspirational, not a hard prerequisite for enrolment:
- Programming experience in any language. C++ experience is helpful but not required.
- Familiarity with the Linux or macOS command-line environment.
- Basic scripting and file manipulation.
- Networking fundamentals (IP addresses, ports, UDP/TCP concepts).
- Version control basics (git clone, commit, push).
Why attend
Capability, not credentials. The programme certificate documents that your engineer has programmed, deployed, and field-operated autonomous surface vehicles under MIT instruction. Mission logs, code, and MOOS-IvP proficiency come home with them. The MOOS-IvP skills and autonomy engineering experience are platform-transferable — applicable to UUV programmes and larger autonomous marine platforms.
Substantive material for proposals. Organizations submitting HCDI 2026 bids or EDF proposals can reference MIT-certified marine autonomy training by named engineers, with documented in-water mission results.
Requirements and preparation
- Prerequisites
- Programming experience in any language. C++ experience helpful but not required. Familiarity with the Linux or macOS command-line environment expected. No prior experience with marine robotics, autonomous vehicles, or MOOS-IvP is assumed — the course starts from fundamentals.
- Laptop
- Linux or macOS required. Detailed software installation instructions will be provided before the course.
- Internet access
- Reliable internet access is required for downloading course materials and lab exercises.
- Software
- MOOS-IvP and associated development tools. Installation guides sent on enrolment confirmation.
- Hardware provided
- All field hardware — including Blue Robotics BlueBoat USVs and payload autonomy computers — is provided. Participants do not need to bring any hardware other than their laptop.
- Pre-course modules
- Enrolled participants receive access to online preparatory modules approximately three weeks before the course begins. These modules cover Linux command-line fundamentals, introductory C++, and an orientation to the MOOS-IvP environment. Completion is strongly recommended, particularly for participants with limited Linux or C++ experience. (Content and availability subject to final confirmation with MIT.)
Programme details
| Dates | June 15–26, 2026 |
| Location | Hellenic Naval Academy, Piraeus, Greece |
| Format | Full-time in-person intensive, Monday–Friday |
| Duration | 2 weeks (10 days of instruction and in-water sessions) |
| Language | English |
| Class size | 30 participants maximum |
| Lead instructor | Michael Benjamin (MIT Marine Autonomy Lab) |
| Certificate | MIT Open Learning certificate on completion |
| Organiser | StartSmart SEE, MIT authorised facilitator in Greece |
Instructor
Supported by MIT teaching assistants from the Marine Autonomy Lab and MIT Lincoln Laboratory.