Repository: https://github.com/isaac-for-healthcare/i4h-workflows

Robotic Ultrasound Docker Container

This guide provides instructions for running robotic ultrasound simulations using Docker containers with Isaac Sim.

Prerequisites

• Docker Engine
• NVIDIA Docker Runtime
• X11 forwarding support (for GUI mode)
• RTI License
• Please refer to the Environment Setup for instructions to prepare the I4H assets and RTI license locally.
• The license file rti_license.dat should be saved in a directory in your host file system, (e.g. ~/docker/rti), which can be mounted to the docker container.

Build the Docker Image

# Clone the repository
git clone https://github.com/isaac-for-healthcare/i4h-workflows.git
cd i4h-workflows
# Build the Docker image with no cache to ensure fresh build
docker build --no-cache -f workflows/robotic_ultrasound/docker/Dockerfile -t robotic_us:latest .

Running the Container

# Allow Docker to access X11 display
xhost +local:docker

# Run container with GUI support
docker run --name isaac-sim -it --gpus all --rm \
    --network=host \
    --runtime=nvidia \
    --entrypoint=bash \
    -e DISPLAY=$DISPLAY \
    -e "OMNI_KIT_ACCEPT_EULA=Y" \
    -e "ACCEPT_EULA=Y" \
    -e "PRIVACY_CONSENT=Y" \
    -v /tmp/.X11-unix:/tmp/.X11-unix \
    -v ~/docker/isaac-sim/cache/kit:/isaac-sim/kit/cache:rw \
    -v ~/docker/isaac-sim/cache/ov:/root/.cache/ov:rw \
    -v ~/docker/isaac-sim/cache/pip:/root/.cache/pip:rw \
    -v ~/docker/isaac-sim/cache/glcache:/root/.cache/nvidia/GLCache:rw \
    -v ~/docker/isaac-sim/cache/computecache:/root/.nv/ComputeCache:rw \
    -v ~/docker/isaac-sim/logs:/root/.nvidia-omniverse/logs:rw \
    -v ~/docker/isaac-sim/data:/root/.local/share/ov/data:rw \
    -v ~/docker/isaac-sim/documents:/root/Documents:rw \
    -v ~/.cache/i4h-assets:/root/.cache/i4h-assets:rw \
    -v ~/.cache/huggingface:/root/.cache/huggingface:rw \
    -v ~/docker/rti:/root/rti:ro \
    robotic_us:latest

Running the Simulation

The command to run the simulation is the same as Running Workflows section.

For example,

# Inside the container
conda activate robotic_ultrasound

# Run simulation with GUI
(python -m policy.run_policy --policy pi0 & python -m simulation.environments.sim_with_dds --enable_cameras & wait)

Troubleshooting

GPU Device Errors

• "Failed to create any GPU devices" or "omni.gpu_foundation_factory.plugin" errors: This indicates GPU device access issues. Try these fixes in order:

Verify NVIDIA drivers and container toolkit installation:

 ```bash
 # Check NVIDIA driver
 nvidia-smi

 # Check Docker can access GPU
 docker run --rm --gpus all --runtime=nvidia nvidia/cuda:12.8.1-devel-ubuntu24.04 nvidia-smi
 ```

If the --runtime=nvidia is not working, you can try to configure Docker daemon for NVIDIA runtime. The file should contain the following content:

 ```json
  {
     "default-runtime": "nvidia",
     "runtimes": {
        "nvidia": {
              "path": "nvidia-container-runtime",
              "runtimeArgs": []
        }
     }
  }
 ```

• Policy not responding: Ensure the policy runner is started before the simulation and is running in the background

• No ultrasound images: Verify that the ultrasound raytracing simulator is running

• Display issues: Make sure xhost +local:docker was run before starting the container and the terminal shouldn't be running in a headless mode (e.g. in ssh connection without -X option)

• Missing assets: Verify that the I4H assets and RTI license are properly mounted and accessible

Verification Commands

After applying fixes, test with these commands:

# Test basic GPU access
docker run --rm --gpus all nvidia/cuda:12.0-base-ubuntu20.04 nvidia-smi

# Test Vulkan support
docker run --rm --gpus all -v /tmp/.X11-unix:/tmp/.X11-unix -e DISPLAY=$DISPLAY robotic_us:latest vulkaninfo

# Test OpenGL support
docker run --rm --gpus all -v /tmp/.X11-unix:/tmp/.X11-unix -e DISPLAY=$DISPLAY robotic_us:latest glxinfo | head -20