Install curobo_ros

This guide takes you from a fresh machine to a running GPU-accelerated trajectory in RViz. Everything runs inside a Docker image so your host stays clean.

Choose Your Workflow

Workflow

Best For

Image Size

curobo_ros

DEV Mode

Modifying curobo_ros internals

~25-30 GB

Mounted from host (live editing)

PROD Mode

Using curobo_ros in your projects

~15-20 GB

Pre-installed in container

Choose DEV if you want to contribute to or modify curobo_ros source code. Choose PROD if you want to use curobo_ros as a dependency in your robot project.

Prerequisites

Requirement

Notes

Ubuntu 20.04 / 22.04 (or Windows 11 + WSL 2)

Host system tested in CI.

NVIDIA GPU with recent drivers

CUDA 12 is used in the container. Check with nvidia-smi.

~30 GB free disk space

Docker image and dependencies are large.

Docker ≥ 24 & NVIDIA Container Toolkit

Follow nvidia.github.io/nvidia-docker.

 
# Verify prerequisites
nvidia-smi
docker --version
docker run --rm --gpus all nvidia/cuda:12.1.0-base-ubuntu22.04 nvidia-smi

PROD Mode Setup

1. Clone & Build Image

cd ~
git clone https://github.com/Lab-CORO/curobo_ros.git --recurse-submodules
cd curobo_ros/docker
bash build_docker.sh  # GPU → PROD (mode 2)

Result: curobo_ros:ampere-prod — smaller image with curobo_ros pre-installed.

2. Create Your Workspace

Create your workspace on the host machine, not inside the container. Docker containers are ephemeral: if you delete or recreate a container, everything stored inside that is not in a mounted volume is lost. By keeping your workspace on the host, your code and built packages are safe regardless of what happens to the container.

# On the host machine
mkdir -p ~/my_robot_ws/src
# Place your packages in ~/my_robot_ws/src/

This path will be requested when you run start_docker_x86.sh (Step 3/3 prompt: Workspace path:). The script mounts it into the container at /home/ros2_ws, so any file you edit on the host is immediately visible inside the container, and vice versa.

3. Start the Container

bash start_docker_x86.sh  # GPU → PROD → enter your workspace path

Your workspace is mounted at /home/ros2_ws. curobo_ros is pre-installed in /home/curobo_ws and auto-sourced on startup.

Daily Workflow

Opening Additional Terminals

Option 1 — VSCode (recommended):

  1. Install the “Dev Containers” extension

  2. Click >< in the bottom-left → “Attach to Running Container…”

  3. Open terminals with Terminal → New Terminal (all run inside the container)

Option 2 — docker exec:

# On your host
docker exec -it curobo_ampere_dev bash
source /home/ros2_ws/install/setup.bash

Starting a Stopped Container

After the first run, never use the start script again — use:

docker start curobo_ampere_dev
docker exec -it curobo_ampere_dev bash

DEV Mode Setup

1. Clone the Repository

mkdir -p ~/ros2_ws/src
cd ~/ros2_ws/src
git clone https://github.com/Lab-CORO/curobo_ros.git --recurse-submodules

2. Import Dependencies

cd ~/ros2_ws/src
sudo apt install python3-vcstool
vcs import < curobo_ros/my.repos

This imports curobo_msgs (custom messages) and curobo_rviz (RViz plugin).

Resulting workspace:

~/ros2_ws/
└── src/
    ├── curobo_ros/
    ├── curobo_msgs/
    └── curobo_rviz/

3. Build the Docker Image

⚠️ Takes 20-30 minutes and requires ~30 GB disk space.

cd ~/ros2_ws/src/curobo_ros/docker
bash build_docker.sh

Interactive prompts:

  1. GPU architecture — Ampere (RTX 30XX) / Ada Lovelace (RTX 40XX) / Turing (RTX 20XX) / Volta

  2. Mode — select 1 for DEV

Result: image named curobo_ros:ampere-dev (or your GPU + mode).

4. Start the Container

⚠️ Run this script once to create the container. Afterward use docker start + docker exec.

bash start_docker_x86.sh

Prompts: same GPU → DEV mode → confirm mount.

What gets mounted:

  • ~/ros2_ws/src/curobo_ros → /home/ros2_ws/src/curobo_ros

  • ~/ros2_ws/src/curobo_rviz → /home/ros2_ws/src/curobo_rviz

  • ~/ros2_ws/src/curobo_msgs → /home/ros2_ws/src/curobo_msgs

GPU access and X11 forwarding (RViz) are enabled automatically.

5. Build the Workspace

All commands below run inside the container.

cd /home/ros2_ws
colcon build

Build time: ~2-5 minutes on first build.

6. Source the Environment

source /opt/ros/humble/setup.bash
source /home/ros2_ws/install/setup.bash

7. Launch the System

ros2 launch curobo_ros gen_traj.launch.py

First launch warmup (~30-60 s): the node compiles GPU kernels. RViz then opens showing the default robot (Doosan M1013).

⚠️ RViz doesn’t appear? See Troubleshooting.

Next Steps