Getting Started

Introduction

The steps from installation to your first policy.

1. Install the Incar system

python -m venv incar
source incar/bin/activate
pip install (according to instructions sent via email)

2. Start the Incar system

incar run          # Starts the incar system

3. Setup your robot

Incar provides an easy to use interface to control your robot and stream robot state.

Integrate your robot with minimal code (see example) !

Camera's can easily be configured, and can furthermore be easily integrated with their native SDK.

4. Connect the headset and robot

All you need to do is to point the headset's cameras to the QR code generated on the GUI. Make sure your headset is on the same network.

5. Teleoperate and collect your data

6. Train and run your policy

With simple commands and a clear GUI interface:

incar train --config_path my_policy.json

Documentation

Check out our documentation for next steps.

FAQ

What Operating System do I need?

The system is tested on Ubuntu 22, Ubuntu 24, and Windows 11. Other operating systems might work as well.

Do you support ROS?

The Incar system does not put any requirements on the robot implementation. So yes, you can use ROS.

Really any robot?

As long as your robot can be controlled with cartesian velocity commands, yes! For hands, joint remapping might be needed based on the hardware of the hand. If your robot only contains cartesian position or joint controllers, you will need to either integrate the velocity commands or compute the IK solution respectively.

I'm stuck -- what now?

Please check the tutorials and documentation, but if that is not sufficient, make an issue on GitHub or contact us by email and we will help you out!

Contact