- 【Seamless Hugging Face Integration】The SO-ARM101, an embodied intelligent robotic arm built on the Hugging Face Lerobot project, seamlessly connects to the ecosystem, providing easy access to shared code, templates, and pre-trained models.
- 【End-to-End Imitation Learning】Using the Hugging Face AI framework, the arm collects data through teleoperation for model training, enabling autonomous task execution and environmental adaptation.
- 【Dual-Camera Vision System】Equipped with both a gripper-mounted camera and an external camera, the system supports both precise manipulation and environmental awareness for accurate imitation learning.
- 【High-Performance Servos】Featuring six 30KG high-torque servos with magnetic feedback, the arm delivers smooth, stable motion, eliminating issues like power deficiency and jitter.
- 【Professional Visual PC Software】Integrated with servo scanning, status monitoring, and trajectory control, the BusLinker V3.0 debugging board simplifies device control and debugging.
Product Description
The SO-ARM101 is an open-source, 3D-printed robotic arm designed for embodied intelligence research. Hardware optimizations address issues like servo power and structural jams, enabling a focus on AI algorithm development. Unlike standard robotic arms, the SO-ARM101 is designed with learning at its core. It can be trained using techniques like imitation learning and reinforcement learning to perform tasks such as object manipulation and motion replication. The entire system—from hardware to software and algorithms—is fully open-source.
① What is the Hugging Face LeRobot?
Hugging Face LeRobot is an open-source robotic ecosystem that includes a Python library and model-sharing platform. It offers tools, datasets, and pre-trained models to help developers, students, and enthusiasts apply technologies like imitation learning and reinforcement learning in robot development.
② Why are they combined?
The SO-ARM101 is an open-source robotic arm, and LeRobot is an AI ecosystem for controlling robots. Together, they provide an integrated platform that combines hardware and AI to make robotics more accessible and powerful.
1. Technical Features
1) End-to-End Imitation Learning
The robotic arm integrates the Hugging Face Al framework, enabling data collection through leader-follower teleoperation and one-click training of imitation learning models. This allows the robotic arm to autonomously perform complex tasks with environmental adaptability.
① Leader-Follower Collaboration & Data Collection
Through a leader-follower architecture, the robotic arm can be remotely controlled to perform various complex tasks. The system simultaneously collects high-quality multimodal operation data, providing a foundation for subsequent intelligent learning.
② Imitation Learning & Model Training
With Hugging Face's machine learning framework, data can be uploaded and models trained with asingle click. With a few simple steps, demonstration actions can be converted into imitation learning models that the Al can process, enabling the arm to rapidly acquire new skills.
③ Strategy Deployment & Autonomous Execution
Trained models can be deployed to the robotic arm with a single click, enabling the automatic execution of complex actions and significantly improving data collection efficiency and operational capability.
2) Dual Vision Collaboration for Global Coverage
The robotic arm is equipped with both a primary and a tertiary camera, enabling dual perception for precise close-range manipulation + global environmental awareness. This allows for millimeter-level precision in control while also understanding complex surroundings, providing strong support for imitation learning.
① Primary Vision Hand-Eye Coordination
The robotic arm is equipped with an accurate vision system that captures the precise position and surface texture of targets at close range. Its efficient feature recognition ability significantly enhances the efficiency of fine-grained data collection.
② Tertiary Vision Global Awareness
The robotic arm can scan and construct an environmental model in real-time, quickly responding to external changes, ensuring precise and smooth execution,and improving task stability and reliability.
2. Why Choose Us?
Not just open-source -but optimized and enhanced!
The original sO-ARM 101 has been extensively optimized with hardware improvements, addressing commonissues in the original design, all while fully adhering to the open-source license.
1) High-torque Magnetic Encoder Bus servos
The robotic arm is equipped with six 30KG high-torque magnetic encoder self-feedback bus servos. Combined with a rotating base and flexible joint movements, the arm achieves smooth and fluid motion, eliminating the power limitations and vibrations found in the original design.
① Upgraded Torque, Precision, and Responsiveness
The servos are powered by a 12V high-voltage supply, providing 30kg.cm of torque with low current draw. This design balances high efficiency and compact size, enabling the robotic arm to handle heavier loads and move more flexibly.
② 12V High Voltage for Improved Power Density
The servos feature 360 magnetic encoder feedbackand an acceleration/deceleration control algorithm, boosting torque, precision, and response speed. This design reduces gear backlash and motion dead zones, resulting in smoother power delivery and more accurate positioning.
③ Different Gear Ratios
The intellient voltage and drive design minimizes heat and power loss.The main arm features a1:147 gear ratio for fast response and fine control, while the follower arm uses a 1:345 high reduction ratio for enhanced torque and better positioning stability.
④ Professional Visual PC Software
The Hiwonder BusLinker V3.0 servo debugging board comes with custom PC software, offering features like servo scanning, parameter setup, status monitoring, and trajectory control. It enables easy adjustments to motion and safety settings for various development needs.
2) Optimized Structural Design
The structure has been fully optimized for higher precision, smoother movement, and greater durability, elevating the intelligent operation experience and eliminating looseness and lag.
① Precision Optimization
Critical joints and load-bearing areas have been re-engineered and reinforced, reducing movement gaps for greater precision and more reliable positioning.
② Smooth Movement Enhancement
The wiring layout and joint design have been optimized to eliminate potential interference and lag, ensuring smooth and seamless motion on every motion.
③ Durability Upgrade
High-strength PLA+/PETG materials are used, with key areas strengthened and thickened to improve fatigue resistance and prevent deformation, ensuring long-term durability.
3. Applications
1) Academic Research
Ideal for teaching experiments and research in fields such as robotics, artificial intelligence, and automation.
2) Developers/Tech Enthusiasts
Used for algorithm development, product prototyping, and participating in robotics competitions.
3) Educational Institutions
An ideal teaching tool for advanced courses like AI programming.
4) Experienced Hobbyists
Perfect for DIY projects, building a cutting-edge robot to practice AI technology.
4. High-Performance Hardware
1) 30KG Magnetic Encoder Self-Feedback Bus Servo
The servo integrates two high-quality magnetic encoder motors with a contactless magnetic feedback design, enabling millisecond-level coordinated response, ensuring every movement of the robotic arm is smooth and fluid.
2) HD Wide-Angle Camera
Equipped with a high-performance light-sensitive chip and wide-angle lens, it delivers precise image capture and processing. Supports manual focus, plug-and-play without the need for drivers.
3) BusLinker V3.0 Servo Debugging Board
The BusLinker V3.0 servo debugging board comes with built-in bus servos, USB, and other communication interfaces. It supports easy expansion for custom development.Simply connect via USB to quickly control servos, offering both simplicity and efficiency.
