OASIS – ROS 2 based Smart Home operating system integrates with Kodi

OASIS is a Smart Home operating system based on ROS 2 that currently implements computer vision, input streaming, and general automation features, and can be integrated into Kodi media center.

The operating system was recently released by Garrett Brown (a.k.a. garbear or eigendude), who is also known for being the RetroPlayer developer from Team Kodi/XBMC, and provides a complete implementation of the Firmata protocol for communicating with Arduino boards, plus additional support for temperature and humidity sensors, I2C, servos, sonar, SPI, stepper motors, and 4-wire CPU fans.

Two main use cases are computer vision and input streaming at this time. The illustration above shows the former with the Kinect 2 driver ported to ROS 2, a background subtractor on all camera feeds using bgslibrary C++ background subtraction library, and Kodi as the visual interface. The second, input streaming, can be seen below with a Lego train (including a Falcon spaceship!) controlled with a PS4 controller through Kodi running on a Raspberry Pi SBC and controlling an Arduino connected to a robotics motor controller that drives the train’s 9V motors.

You’ll find the source code for OASIS ROS 2-based OS on Github, and Garrett also submitted two pull requests to the Kodi repo to add support for computer vision-based interaction for Smart Home control.

We learn a few more interesting tidbits from the commits’ descriptions:

The approach I’ve taken with smart home is a bit unique: I built on ROS 2. While this introduces some heavy overhead, building on an industrial decentralized communication framework allows for scalability to virtually unlimited low power smart home devices. My smart home’s computation graph is currently at 8 Linux nodes and 3 Arduino nodes and growing.

… I’m not seriously proposing we merge this and add ROS 2 as a dependency; it adds 2 millions lines of code. I’m just sharing the code I run everyday at home, and maybe it’ll inspire someone.

So it would be quite a large dependency, but may not be integrated into Kodi by default, but at least, if you intend to add computer vision and smart home controls to Kodi, the code is there. He also appears to have used this implementation for a while saying the “computer vision pipeline has been running relatively stable for about 10 months now”.

Thanks to Hedda for the tip

Support CNX Software! Donate via cryptocurrencies, become a Patron on Patreon, or purchase goods on Amazon or Aliexpress