ax-ros2-ref is an experimental, unsupported reference implementation that integrates the high-performance, low-power Axelera Metis AIPU (214 TOPS, 5-9W) with ROS2 Jazzy, enabling real-time object detection and pick-and-place robotics.

Stars

5

7-day growth

No data

Forks

0

Open issues

0

License

MIT

Last updated

2026-07-28

AI repository intelligence
FR-AI / ANALYSIS

Why it is worth attention

It demonstrates edge AI inference at a remarkably high TOPS-per-watt ratio (214 TOPS at 5-9W) directly within the ROS2 ecosystem, with out-of-the-box compatibility with Nav2, MoveIt2, and Foxglove, and includes a complete pick-and-place demo using a Niryo NED2 arm.

Who it is for

  • ROS2 robotics developers
  • Edge AI and embedded vision engineers
  • Robotics researchers evaluating Axelera AIPUs
  • Hobbyists and makers building robot stacks with high-performance inference

Use cases

  • Real-time object detection in ROS2-based robots
  • Pick-and-place automation with a robotic arm
  • Integrating edge AI accelerators into existing ROS2 pipelines
  • Prototyping and evaluating Axelera hardware for robotics

Strengths

  • Extremely high TOPS per watt (214 TOPS at 5-9W) compared to Jetson Orin Nano Super (67 TOPS at 15-25W)
  • Quick three-command setup from cloning to launch
  • Works out of the box with Nav2, MoveIt2, and Foxglove
  • Includes a full 14-state pick-and-place demo with MoveIt2 integration

Considerations

  • Experimental and unsupported – no warranty, no maintenance guarantees, APIs may break
  • Only tested with ROS2 Jazzy and Voyager SDK versions 1.5.3-1 and 1.6.0-1; other versions may not work
  • Requires specific Axelera hardware (Metis M.2) and the Voyager SDK, limiting adoption to users with that hardware

README quick start

Quick start (3 commands)

Description

Experimental reference implementation of Metis-accelerated inference for ROS2

Related repositories

Similar projects matched by category, topics, and programming language.

lopopolo
Featured
lopopolo GitHub avatar

harness-engineering

Harness Engineering is a methodology for improving coding agent outputs by carefully crafting the environment around them—providing curated context, tools, and executable constraints that encode an organization’s nonfunctional requirements and cumulative lessons.

AI & Machine LearningAI Agents
2,390
slvDev
Featured
slvDev GitHub avatar

esp32-ai

A 28.9 million parameter language model runs on an $8 ESP32-S3 microcontroller entirely on-device, generating simple stories at about 9.5 tokens per second.

AI & Machine LearningLarge Language Models
1,960
littledivy
Featured
littledivy GitHub avatar

mimic

mimic captures traffic from any iOS or web app and automatically generates a Python client library that lets you call the app's API like a regular library.

AI & Machine Learning
1,482