Blog
StarArm 102 Officially Featured in Hugging Face LeRobot Documentation
StarArm 102 has been officially included in the Hugging Face LeRobot documentation, becoming one of the officially supported robotic hardware platforms. This inclusion enables developers to directly utilize StarArm 102 following LeRobot's standardized workflow for device configuration, teleoperation, data collection, model training, and deployment—eliminating the need for complex adaptation development.
📄 Official Documentation: https://huggingface.co/docs/lerobot/rebot_b601
Background and Significance
In recent years, open-source robotics frameworks represented by LeRobot have emerged as critical development platforms in the fields of embodied intelligence, robot learning, and imitation learning. An increasing number of universities, research institutions, and development teams are leveraging this ecosystem to conduct algorithm research, data collection, and robotics application validation.
🎯 The official inclusion of StarArm 102 in the LeRobot documentation represents a significant milestone: the product has achieved standardized integration with mainstream international development ecosystems. Developers can now access the device by following the official documentation and utilize LeRobot's standard workflow for data collection, behavior cloning training, and robot control—without the need to develop custom adaptation interfaces.
StarArm 102: Open-Source Robotic Arm Platform Overview
StarArm 102 is a 6+1 degree-of-freedom (DOF) open-source robotic arm teleoperation system available in three product configurations:
Product Configurations
🔹 LD (Lightweight)
Ultra-lightweight handheld design
Weight: 663g
🔹 HD (High-torque)
Professional-grade hovering functionality
Servo: RP8-U45H-M
🔹 FL (Follower Arm)
Execution arm
Precise trajectory reproduction
Six Core Product Features
1. Ultra-Lightweight Handheld Design
With a total weight of only 663g, the StarArm 102-LD ranks among the lightest master arm devices in its class. Operators experience no hand fatigue even during continuous sessions exceeding 30 minutes.
2. High-Precision Trajectory Acquisition
All seven joints are equipped with 12-bit magnetic encoders, providing an angular resolution of 0.088°.
3. Cross-Brand Compatibility
Beyond native StarArm 102-FL, the system directly supports third-party arms including reBot B601-DM and Xinghai Tu platforms.
4. Plug-and-Play Deployment
Requires only a 12V 3A power supply. USB serial direct connection enables driver-free operation on Ubuntu 22.04. Total deployment time: under 5 minutes.
5. Four Development Methodologies
| Method | Difficulty | Capabilities |
|---|---|---|
| Web-Based Visual Control | Zero-code | Browser-based teleoperation |
| Python SDK | Easy | Single line to start |
| ROS 2 | Medium | MoveIt2 integration |
| LeRobot (Official Certification) | Advanced | Hugging Face certified |
🏆 LeRobot Official Certification Advantages:
• Direct acquisition of HuggingFace-compatible datasets
• Zero secondary adaptation required
• Cutting-edge embodied AI research ready
6. Upgradeable HD Configuration
The HD model features hovering functionality with RP8-U45H-M high-torque servos. The arm maintains any pose when released. Operational reach: 420mm.
Application Scenarios
🎓 University Education
• Robotics curriculum
• Student projects
• Graduation design
🔬 Scientific Research
• Imitation learning
• Teleoperation research
• Embodied intelligence
🤖 Robotics Algorithm
• Motion planning
• Control testing
• Reinforcement learning
🧠 Embodied AI
• AI + robotics integration
• Multimodal learning
Technical Resources
- 📄 Official LeRobot Documentation: https://huggingface.co/docs/lerobot/rebot_b601
- 📚 FashionStar Product Wiki: https://fashionstar.com.hk/wiki/
- 🏢 Corporate Website: https://fashionstar.com.hk/
- 🔧 Hardware Assembly Guide: https://wiki.seeedstudio.com/rebot_arm_b601_dm_lerobot/
