ASPEN lets researchers place digital twins of real underwater vehicles into scientifically accurate simulations of real ocean locations. The vehicles behave like their physical counterparts — they follow the same physics, carry the same sensors, and navigate the same currents. Missions can be planned, tested, and analyzed entirely in software before any hardware goes in the water.
ASPEN was developed under U.S. government contract. Software rights are governed by DFARS 252.227-7014 (Rights in Other Than Commercial Computer Software and Other Than Commercial Computer Software Documentation). NPS has received delivery of the technical data package and associated rights as specified under this clause.
Core simulation engine and orchestration layer. ASPENSDK is the central repository that ties everything together — it manages simulation lifecycle, coordinates vehicle spawning, loads environments, and provides the Python API surface that researchers interact with. Runs on top of NVIDIA Isaac Sim.
Processes real-world ocean data from sources like NOAA and Navy oceanographic models into simulation-ready formats. Handles bathymetry, ocean currents, temperature, salinity, and sound speed profiles — converting raw scientific datasets into volumetric representations that Isaac Sim can render and query at runtime.
Six-degree-of-freedom (6-DOF) physics models for 4 real underwater vehicle platforms, 11 simulated sensors, PID control systems, and ROS 2 integration. Each vehicle model replicates the hydrodynamics, buoyancy, and control characteristics of its real-world counterpart.
9.5 GB of 3D models (USD format), pre-processed ocean environment data, and simulation-ready scene files for 5 real-world locations. These are the ready-to-load assets that let researchers start simulating immediately without processing raw data from scratch.
Browser-based Command and Control interface built with React and TypeScript. Provides mission planning, real-time vehicle monitoring, waypoint management, and 2D map visualization. Connects to ASPENSDK via WebSocket to control running simulations without needing to touch Python code or Isaac Sim directly.
Simulate autonomous underwater navigation with physically accurate 6-DOF vehicle dynamics, buoyancy models, and PID control systems that match real vehicle behavior.
Model ocean environments from real NOAA and Navy data — bathymetry, currents, temperature, salinity, and sound speed profiles for actual geographic locations.
Run coordinated multi-vehicle missions with multiple AUVs operating simultaneously, enabling fleet coordination and swarm behavior research.
Compute underwater acoustic propagation using Bellhop ray tracing, accounting for sound speed profiles, bathymetry, and environmental conditions that affect sonar performance.
Train AI agents via reinforcement learning using GPU-accelerated simulation. The environment provides the OpenAI Gym-style interface needed for RL algorithm integration.
Connect simulated vehicles to real autonomy software via ROS 2. The same control code that runs in simulation can run on physical vehicles with minimal changes.
Visualize missions in full 3D through NVIDIA Omniverse with RTX rendering, or through the browser-based C2 UI for 2D map-based mission monitoring and control.
The team members listed above are sourced from contributor records across the ASPEN repositories. Role descriptions are approximate and based on contribution patterns. For the authoritative contributor list, see the CONTRIBUTORS.md file in each repository.
NPS is receiving the first NVIDIA DGX GB300 NVL72 in the Department of the Navy — a system purpose-built for the kind of GPU-accelerated simulation and AI training that ASPEN demands. With 72 Blackwell Ultra GPUs and over 20 TB of GPU memory, this machine transforms what is possible with ASPEN.
What was previously limited to single-vehicle simulations on a workstation can now scale to fleet-level operations, Monte Carlo analysis across thousands of runs, and reinforcement learning campaigns that would be impractical on any other system in the Navy.
ASPEN on the GB300 is not limited to a single department or lab. It is a shared research platform available to 1,500+ students across NPS, enabling work in autonomous systems, oceanography, operations research, computer science, and defense analysis. Any researcher studying undersea warfare, autonomous navigation, or ocean sensing can use this platform to test hypotheses at a scale that was previously impossible.