Project Logs

Collapse

• OpenAMRobot - First Public Release v0.1

  Alex Reznichenko • 07/09/2026 at 10:42 • 0 comments

  # OpenAMRobot

  Using a robot shouldn't require a robotics team.

  **OpenAMRobot** is an open-source platform built around a simple idea: the person who knows the job should be able to teach a robot how to do it - by showing it, not by writing code.

  Our long-term vision is **"Show → Correct → Judge."**

  An operator demonstrates a task, intervenes when necessary, and labels successful or unsuccessful attempts. Behind the scenes, this becomes imitation learning, reward modeling, and reinforcement learning - but the user never has to deal with AI pipelines or robotics programming. Domain expertise becomes the programming language.

  ## Demo

  🎥

  ---

  # 🚀 First Public Release

  This first release focuses on a complete **open-source autonomous mobile robot (AMR) platform** with separated Hardware, Software, and Firmware repositories.

  ## Hardware

  - Differential-drive mobile robot
  - Laser-cut and bent sheet-metal chassis
  - Raspberry Pi 5 + Teensy 4.0 architecture
  - LiDAR, IMU, wheel encoders
  - Complete CAD models
  - Manufacturing documentation
  - Wiring diagrams
  - Bill of Materials (BOM)
  - Datasheets and supplier references
  - Electrical documentation
  - Safety documentation

  ## Software (ROS 2 Jazzy)

  - ROS 2 Jazzy
  - Gazebo Harmonic simulation
  - Nav2 autonomous navigation
  - LiDAR SLAM
  - AprilTag-based visual auto-docking
  - Visual odometry
  - Robot description (URDF/Xacro)
  - Complete launch system
  - Docker development environment
  - Well-structured ROS 2 packages

  ## Firmware

  - Teensy 4.0 firmware
  - micro-ROS integration
  - Differential drive control
  - Encoder odometry
  - IMU integration
  - PID motor control
  - Debug and telemetry interface
  - Motion safety framework
  - Calibration tools

  ## User Interface

  - Modern web-based interface
  - Blockly no-code programming
  - Live telemetry
  - Interactive maps
  - Camera streaming
  - Robot diagnostics
  - Voice / LLM command interface

  ---

  # 📦 What's Included

  - Complete Gazebo simulation
  - ROS 2 navigation stack
  - Autonomous docking
  - Differential-drive firmware
  - Hardware documentation
  - Mechanical CAD
  - Electrical documentation
  - Firmware
  - Manufacturing files
  - Docker environment
  - Well-structured repositories
  - MIT License
  - Fully open source

  ---

  # 🔭 Roadmap

  The mobile robot is only the beginning.

  The next major milestones include:

  - Industrial-grade modular electronics
  - Custom PCBs
  - Functional safety
  - Wireless charging
  - Dual-arm mobile manipulator
  - Vision-based manipulation
  - Foundation-model fine-tuning (π0 / NVIDIA GR00T)
  - Simulation-based data generation
  - Human-in-the-loop reinforcement learning

  Ultimately, OpenAMRobot aims to become an open platform for autonomous mobile manipulation rather than only a mobile robot.

  ---

  # ❤️ Building in Public

  This project is still early, and we're intentionally building it in the open.

  Several companies are already building robots based on OpenAMRobot, and their feedback is helping us improve both the hardware and software. We welcome bug reports, ideas, pull requests, and discussions from engineers, researchers, makers, and robotics companies.

  If you'd like to contribute - or build your own robot - we'd love to have you join us.

  ---

  ## Repositories

  - 🛠️ Hardware: https://github.com/openAMRobot/openamr-platform-hw
  - 🤖 Software: https://github.com/openAMRobot/openamr-platform-sw
  - ⚙️ Firmware: https://github.com/openAMRobot/openamr-platform-fw

• v0.01 - first open release

  Alex Reznichenko • 07/07/2026 at 22:37 • 0 comments

  The idea: teach a robot by demonstration instead of programming it - show, correct, judge → imitation learning + reward model + RL, with the operator never touching the ML.

  Live today (early, honest): open mobile base with autonomous nav (ROS 2 + Nav2 + RPLIDAR SLAM), no-code Blockly UI with live telemetry + mapping, and an LLM voice-command layer. Two-layer architecture: Raspberry Pi 5 (high-level: ROS 2, SLAM, nav, UI) + Teensy 4.0 (real-time: motors, encoders, IMU). Differential drive, 24V BLDC motors. Documented, MIT-licensed.

  Next: harden electronics (custom PCB replacing the breadboard, safety layer, power distribution), then start the dual-arm manipulation + learn-by-demonstration pipeline (foundation-model fine-tuning + simulation + human-in-the-loop RL). The mobile base is the on-ramp; the learning platform is the real project.

  Rough in places, built in public. Feedback and contributors very welcome.
  Repo: github.com/openAMRobot

View all 2 project logs