ASPEN Demo Playbook

How to deliver an effective ASPEN demonstration

Executive Talking Points

Key Messages for Leadership

  • ASPEN is a digital ocean — a physically accurate simulation of underwater vehicles and their operating environment
  • Built on NVIDIA's latest platform — the same technology powering autonomous cars and digital twins
  • Simulates real vehicles (REMUS, BlueROV2) with real ocean data from NOAA and Navy models
  • Enables research at scale: fleet coordination, AI training, and mission rehearsal
  • NPS will be the first Navy command with DGX GB300 — ASPEN is a flagship use case
  • The portable demo runs on a single laptop; the same code scales to supercomputer for full research campaigns

Pre-Demo Checklist

Hardware & Environment

  • Alienware laptop fully charged, power adapter accessible
  • Ubuntu booted (not Windows)
  • Isaac Sim installed and verified — run nvidia-smi, confirm CUDA version
  • ASPEN scenarios tested within the last 24 hours
  • C2 UI running in Chrome (npm run dev)
  • External display connected and tested (if presenting to a room)

Backup & Timing

  • Screen recording of a successful demo run saved as backup (in case of hardware failure)
  • Plan for 10–15 minutes of demo with 5 minutes of Q&A

The C2 Interface — What It Is and Why It Matters In Development

What Is It?

The ASPEN Command & Control UI is a professional-grade operational interface — a single-page web application built on React 19, TypeScript, Mapbox GL, and deck.gl. It was designed from the ground up to look, feel, and function like the kind of interface a Navy operator would use in a real mission environment. It is not a debug tool or a research dashboard bolted on top of a simulation. It is an operationally realistic front end that demonstrates what full-stack autonomous undersea operations would look like.

The interface runs in any modern browser and connects to the ASPEN simulation backend via a REST API and WebSocket stream. During the demo it runs against a high-fidelity mock backend that returns realistic data, correct data types, and real-time telemetry updates — indistinguishable from a live connection.

Why It Matters for the Demo

Simulation engines are hard for non-engineers to evaluate. Isaac Sim is powerful, but a raw 3D viewport with vehicle meshes and telemetry logs does not communicate operational value to a commander, program manager, or decision-maker. The C2 UI translates the simulation into the language of operations.

When an audience sees a mission being planned on a real nautical chart, watches multiple vehicles execute autonomously, sees a comms degradation alert fire in real time, and then reviews the deviation analysis — they understand what ASPEN enables. The C2 UI is the argument that this platform is operationally relevant, not just technically interesting.

Four Modes — Four Phases of Operations

▶ PLAN

Mission authoring. Place waypoints on the chart, assign vehicles, define operational area and geofence, set per-waypoint behaviors (transit, survey, loiter, hover, surface), configure abort conditions and recovery point. Validates against vehicle capabilities and depth constraints before execution.

▶ SIMULATE

Mission rehearsal. Plays back a physics-generated telemetry series with adjustable speed (1x–16x). Environmental parameters (current strength, sea state, thermocline depth, water temperature, visibility) are tunable in real time — change conditions and regenerate the run to see mission impact. Event log shows waypoint completions, alerts, and autonomy decisions as they occur in sim time.

▶ MONITOR

Live operations. Vehicles update position, heading, and sensor status at 1-second intervals via WebSocket. Battery depletion, comms degradation, and autonomy state changes appear in real time. Fleet health panel gives an at-a-glance status for all assigned vehicles. Alert queue surfaces critical conditions with severity classification.

▶ ANALYZE

Post-mission assessment. Overlays planned track against actual track, marks deviations with autonomy-generated explanations (why did the vehicle deviate, what behavior triggered it, what was the confidence). Coverage heatmap shows sensor footprint versus tasked area. Metrics panel: coverage percentage, distance traveled, deviations logged, estimated mission success probability.

Key Features to Highlight

  • Operational realism — Classification banner (UNCLASSIFIED / SECRET / TOP SECRET) renders on every screen. MIL-STD-2525D symbology for all vehicles and overlays. This is the visual vocabulary of naval operations.
  • Map-centric design — The Mapbox GL / deck.gl viewport is the primary workspace, not a secondary view. All planning, monitoring, and analysis happens in the geographic context where operations occur.
  • Multi-vehicle fleet management — Supports concurrent vehicles with independent telemetry, sensor configurations, and autonomy states. Each vehicle card shows battery, comms health, depth, speed, and current behavior.
  • Environmental awareness — Eight toggleable oceanographic layers: bathymetry, currents, temperature, salinity, sound velocity, wave height, ice coverage, and visibility. Each has configurable warning and critical thresholds that fire alerts during operations.
  • Sound velocity profiling — Real-time SVP chart for any geographic position, showing the thermocline structure that governs acoustic sensor performance. Directly relevant to sonar operations.
  • Mission validation — Before execution, the system validates waypoint count, depth constraint compliance, vehicle assignment, battery feasibility, and communications coverage. Validation results are color-coded by severity.
  • Integration-ready architecture — The entire API surface is defined and contracted. Swapping the mock backend for a live ASPEN simulation requires implementing one TypeScript class. No component changes. No type changes. The interface was built for this.

Demo Flow

1

The Problem

2 min
"Testing autonomous underwater vehicles in the real ocean is expensive, slow, and risky."
  1. Set the context: cost of sea time, vehicle risk, limited test windows
  2. Pose the question:
"What if you could run thousands of missions before ever getting wet?"
2

The C2 Interface In Development

4 min
"This is the command-and-control interface — a browser-based operational dashboard built specifically for ASPEN. It's what an operator would use to plan, run, and analyze a mission."
  1. Open the C2 UI in Chrome — point out the classification banner at the top and bottom. This is a military-grade interface, not a research prototype.
  2. Show the four mode tabs along the top: Plan, Simulate, Monitor, Analyze. Each maps to a real phase of autonomous vehicle operations.
  3. In Plan mode, open the Strait of Hormuz MCM scenario. Walk through the map-centric layout — waypoints on the chart, the vehicle assignment panel on the left, environmental layers on the right.
  4. Click a waypoint and show the behavior selector (transit, survey, loiter, hover, surface) and depth constraint. This is how a mission planner communicates intent to the vehicle autonomy stack.
  5. Switch to Simulate mode and press play. Point out vehicles moving along their planned tracks, the event log firing (waypoint reached, comms degraded, battery warning), and the telemetry timeline at the bottom.
  6. Switch to Monitor mode (Arctic Survey scenario). Vehicles are live — battery draining, comms dropping and restoring, autonomy state updating in real time. This is the view during an active operation.
  7. Switch to Analyze mode (South China Sea ISR). Show the planned vs. actual track overlay and the deviation markers with autonomy explanations. This is how a commander reviews what happened and why.
The C2 UI currently runs on a high-fidelity mock backend. Every data type, API endpoint, and WebSocket message is fully defined — the interface is integration-ready. Connecting it to the live simulation engine is a defined engineering task, not an open research question.
3

The Simulation Engine

4 min
  1. Switch to the Isaac Sim window
  2. Show the 3D rendered ocean with bathymetry and vehicles
  3. Run the Dabob Bay scenario with 2–3 REMUS vehicles
  4. Point out: vehicles reacting to currents, navigating autonomously, sensors querying the ocean
  5. Pause the simulation and show vehicle telemetry
4

The Scale Story

2 min
"This laptop runs 10–20 vehicles in real time."
"On the DGX GB300 arriving at NPS: hundreds of vehicles, thousands of Monte Carlo runs."
"Same code, same platform — just more GPU power."
  1. If the audience is academic, show the thesis topics list
  2. Connect the scale story to their specific research interests
5

Q&A

3–5 min
  1. Open the floor for questions
  2. Refer to the talking points below for common questions
  3. If a question is outside your depth, note it and follow up — never guess

Talking Points for Common Questions

"How accurate is the physics?"
Full 6-DOF hydrodynamic models with 37+ parameters per vehicle. Uses the same PhysX engine as NVIDIA's autonomous vehicle program.
"Is the ocean data real?"
Yes. Sourced from NOAA, Navy NCOM/HYCOM models, and GEBCO global bathymetry. The data is time-varying and includes tidal cycles.
"Can this connect to real vehicles?"
Yes, via ROS 2. The BlueROV2 model includes a ROS bridge specifically designed for sim-to-real transfer.
"What's the relationship with NVIDIA?"
NPS has a CRADA with NVIDIA. ASPEN runs on their Isaac Sim platform. The DGX GB300 arriving at NPS is part of that partnership.
"How long to get a student productive?"
2–4 weeks to complete NVIDIA courses and run first simulations. 1–2 months to modify and extend the platform.
"What other locations can we simulate?"
Any location with bathymetry and ocean data. GEBCO provides global coverage. Adding a new location takes 1–2 days of data processing.

Troubleshooting

Isaac Sim won't start
Check nvidia-smi output. Verify CUDA version matches what Isaac Sim expects.
Black screen in simulation
GPU driver issue. Try sudo apt install nvidia-driver-560 and reboot.
Vehicles don't move
Check that Git LFS files were pulled. Run git lfs pull in the assets/ directory.
C2 UI shows blank map
Mapbox token expired or missing in the .env file. Regenerate from the Mapbox dashboard.
Simulation is slow
Close other GPU applications. Ensure the power adapter is connected — the laptop throttles on battery.
NUCLEAR OPTION: If nothing works, switch to the pre-recorded screen capture backup. A smooth recording beats a broken live demo every time.