Prev-Talk | Next-Talk | All-Talks | Talks-Sorted | MOOS-DAWG'26

Talk-03: Two Years With the MIT Spurdog UUV

Raymond Turrisi, Tyler Paine, Michael Benjamin

MIT Department of Mechanical Engineering, MIT Marine Autonomy Lab

The MIT Spurdog AUV is a low-cost, torpedo-shaped autonomous underwater vehicle with a single thruster and fin-based control, designed to lower the barrier for multi-agent and field robotics research. Capable undersea autonomy has historically been expensive and difficult to reproduce, which limits how many researchers can field novel technology at scale. The Spurdog addresses this gap by offering an open and low-cost platform on which researchers can reduce technical risk before translating their work to larger and more boutique platforms. The core vehicle is assembled from commercially available and 3D-printed components for under $3,500, and similarly a navigation nosecone around $5,000 (including DVL+AHRS). This talk reflects on two years of deploying, developing, testing, and hardening the platform, and how perspectives on the application of such vehicles have evolved over time.

A vehicle assembly typically comprises three parts: a core vehicle, a payload, and a nosecone. The core vehicle consists of the Trident Tailcone and the core electrical assembly. The first is the A-size-ready Trident 3-axis aft section, a modular tailcone with a pluggable servo drive module and swappable fins; it has been fielded across hundreds of deployments on four Spurdog UUVs developed to date. The core electrical assembly is intended to be simple and, when needed, replaceable with more rugged or secure technology; at a minimum it comprises a Blue Robotics battery or a custom AA battery pack, a Raspberry Pi 4 with the commercially available Navigator Flight Controller from Blue Robotics, and a 5 Volt/15 Amp power supply. Payloads and nosecones attach through the Sequential Coupling System: a standardized mechanical, power, and communication interface that lets payloads be integrated and hot-swapped in the field, rather than requiring a purpose-built vehicle for each experiment.

The autonomy stack, moos-ivp-spurdog, is a portable extension of MOOS-IvP that has grown to roughly 67,000 lines of original C++ across 94 components. It includes state estimation, a unified acoustic-modem layer spanning the WHOI Micro-Modem, Ahoi, and Popoto modems and their simulated counterparts, a 6-DOF simulator ported from Fossen's MSS toolbox, and an s3mp mission-planning toolbox that compiles high-level mission descriptions to helm configurations and statically analyzes and validates them, so that configuration errors are caught on the bench rather than as helm failures in the water. The software suite can be added to any AUV; only the single Navigator interface application is specific to the Spurdog.

Field work on the Charles River anchored much of this development. We summarize results in GNSS-denied cooperative navigation, including two simultaneously submerged AUVs performing live inter-vehicle acoustic ranging and reaching roughly 1.3 m of terminal error, and ASV-aided navigation in which surface vehicles serve as mobile acoustic references to bound the drift of a submerged AUV. We also discuss the less visible but equally important ruggedization efforts for field deployments ? ranging from new tools for calibrating and diagnosing magnetic heading references to successful testing to 400 meters depth at WHOI.

The result is a reproducible, field-validated, MOOS-IvP-native AUV that is now source-available under a controlled license to U.S. national labs and affiliated institutions, with its contributions documented across five peer-reviewed publications and an accompanying online partner forum. We close with the path forward for continued development, and some exciting announcements regarding work in the coming year.

1. R. Turrisi, H. Huang, F. Str?mstad, J. Morrison, A. Papalia, J. Leonard, and M. R. Benjamin, "The Spurdog AUV: A Field Configurable and Optionally A-Sized Low-Cost AUV," in Proc. IEEE/OES Autonomous Underwater Vehicles Symp. (AUV), 2024. PDF, Google Scholar

2. R. Turrisi and M. R. Benjamin, "A Feature-Rich and Modular Aft Section for Autonomous Underwater Vehicles: The Trident Series of Tailcones," in Proc. OCEANS 2024 - Halifax, Halifax, NS, Canada, 2024. PDF, Google Scholar

3. R. Turrisi, D. A. Duecker, J. Morrison, F. Steinmetz, and M. R. Benjamin, "ASV-Aided AUV Navigation: A Field Study on Nonlinear Estimation for Localization of Low-Cost, Scalable Systems," in Proc. IEEE/RSJ Int. Conf. Intelligent Robots and Systems (IROS), Hangzhou, China, 2025, pp. 12735?12742. PDF, Google Scholar

4. H. Huang, R. Turrisi, R. Leonard, and M. R. Benjamin, "Characterizing Mathematically Parameterized Toroidal Propellers for AUV Applications," in Proc. OCEANS 2025 - Great Lakes, 2025. PDF, Google Scholar

5. E. Gajski, R. Turrisi, J. Hong, J. Morrison, A. Papalia, E. Gallimore, M. R. Benjamin, and J. Leonard, "A Simulation and Field Study Analysis of Cooperative Range-Aided Localization Strategies for Compact, Low-Cost AUV Teams," in Proc. IEEE/OES Autonomous Underwater Vehicles Symp. (AUV), 2026 (pre-print). PDF

Categories:

  • UUVs