Hardware Requirements
This page lists the hardware you need to follow along with the exercises in each module. Many exercises can be completed in simulation, but physical hardware is needed for the full experience.
Computer Requirements
| Component | Minimum | Recommended |
|---|---|---|
| CPU | Intel i5 / AMD Ryzen 5 | Intel i7 / AMD Ryzen 7 or better |
| RAM | 16 GB | 32 GB |
| GPU | NVIDIA GTX 1060 (6 GB) | NVIDIA RTX 3070 or better |
| Storage | 100 GB free SSD | 250 GB free NVMe SSD |
| OS | Ubuntu 22.04 LTS | Ubuntu 22.04 LTS |
An NVIDIA GPU with CUDA support is required for Modules 3 and 4 (Isaac Sim and VLA training). Modules 1 and 2 can be completed without a GPU, though simulation performance will be limited.
Software Requirements
| Software | Version | Used In |
|---|---|---|
| ROS 2 | Humble Hawksbill | Module 1 |
| Gazebo | Harmonic | Module 2 |
| Unity | 2022.3 LTS + ML-Agents | Module 2 |
| NVIDIA Isaac Sim | 4.0+ | Module 3 |
| Python | 3.10+ | All modules |
| PyTorch | 2.0+ | Modules 3, 4 |
| CUDA Toolkit | 12.0+ | Modules 3, 4 |
Optional Physical Robot Hardware
For the full sim-to-real experience in Modules 2 and 3:
Recommended Robot Kits
| Robot | Price Range | Best For |
|---|---|---|
| TurtleBot 4 | $1,200–$1,900 | Navigation (Module 1–2) |
| Franka Emika Panda | $10,000+ | Manipulation (Module 3) |
| WidowX 250 | $3,500 | Budget manipulation arm |
| Open Manipulator-X | $500 | Entry-level manipulation |
Sensors
| Sensor | Purpose | Modules |
|---|---|---|
| Intel RealSense D435 | Depth camera | 2, 3, 4 |
| RPLidar A1 | 2D LIDAR for navigation | 1, 2 |
| IMU (e.g., BNO055) | Inertial measurement | 1, 2 |
Don't worry — all exercises can be completed in simulation. Physical hardware enhances the learning experience but is not required to complete the course. Modules 1 and 2 provide full Gazebo simulations for every exercise.
Cloud Alternatives
If your local machine doesn't meet GPU requirements:
- NVIDIA NGC — Free Isaac Sim containers on NVIDIA's cloud
- Google Colab Pro — GPU instances for PyTorch training (Modules 3–4)
- AWS RoboMaker — Cloud-based Gazebo simulation (Module 2)
Setup Verification
After installing the required software, run the following to verify your setup:
# Verify ROS 2
ros2 --version
# Verify Gazebo
gz sim --version
# Verify Python and PyTorch
python3 -c "import torch; print(f'PyTorch {torch.__version__}, CUDA: {torch.cuda.is_available()}')"
# Verify NVIDIA drivers
nvidia-smi
If any command fails, refer to the installation guides linked in each module's first chapter.