Avocado OS for Jetson Orin (built on OE4T’s meta-tegra BSP) - feedback welcome

Robotics & Edge Computing Jetson Systems Jetson AGX Orin
, , , , , , ,
1

Hi all,

I wanted to share how we’re approaching Jetson support with Avocado OS and get feedback from others building with Yocto on Jetson.

OE4T’s meta-tegra
As many here know, the OE4T project maintains meta-tegra, a Yocto-compatible BSP layer for Jetson platforms. It builds on NVIDIA’s L4T foundation to provide kernel, bootloader, and board enablement for devices like Orin NX, Orin Nano, and AGX Orin.

Avocado OS
We’ve built Avocado OS on top of that, providing a turnkey runtime that integrates meta-tegra out of the box and extends it with:

  • Minimal kernel baseline with optional kernel modules for cameras, sensors, and peripherals

  • Pre-built packages and solution stacks, enabled via systemd / systemd-sysext

  • Support for Basler camera modules, GStreamer, OpenCV, ROS 2 Humble, and Python packages geared toward AI, vision, and robotics

  • Out-of-the-box CI/CD and OTA pipeline support

  • A focus on reproducibility and long-term maintainability

We’re currently targeting Jetson Orin NX / AGX Orin and are watching for the next meta-tegra release for Jetson Thor to expand support.

We recently collaborated with Advantech on the iCAM-540 and are working to introduce additional Jetson-based solutions with pre-integrated tools, integrations, and libraries that this community commonly uses.

If there’s a package, stack, or module missing that would accelerate your project, we’d love to hear about it — for those of you building on Yocto, or considering alternatives to JetPack:

  • What stacks and modules do you most often integrate?

  • How closely do you track upstream meta-tegra, and where have you had to fill in the gaps yourself?

Looking forward to your input.

6

Thanks for your sharing to the community!

1 Like