6DoF HiWonder LeArm Project Update to try and fix (WIP)
📌 Last updated: 18/02/2025
🔗 GitHub Organization & Repo: 6DoF-HiWonder-LeArm
This project documents my 6DoF robotic arm setup using the HiWonder LeArm, ROS2, and AI-based automation. This blog is a work in progress (WIP), and I will continue to update it as I make advancements.
🛠️ Creating CAD & URDF for ROS2 Simulation
Since no official CAD files existed for the HiWonder LeArm, I had to manually create 3D models for each component. This step was essential to generate a URDF file, which enables simulation in RViz2 under ROS2.
Challenges Faced:
✔ Reverse-engineering dimensions from existing hardware.
✔ Ensuring correct joint limits & transformations in URDF.
✔ Aligning CAD models with real-world physical constraints.
After completing the URDF model, I was able to visualize and simulate the robotic arm in RViz2, making it easier to plan kinematics before moving to real-world testing.
🔌 Communicating with LSC-6 via UART (Update: 18/02/2025)
In my latest update, I successfully established communication between the Raspberry Pi 5 and the LSC-6 servo controller using UART and a logic level converter.
Key Achievements:
✔ Configured /dev/ttyAMA0 for serial communication.
✔ Sent and received positioning commands for each joint.
✔ Verified servo angles through real-time feedback.
This marks a major milestone, as I can now fully control the arm through ROS2 instead of relying on HiWonder’s built-in software.
🤖 Kinematics & AI-Powered Welding (Ongoing Work)
Current Progress:
- ✅ Implemented kinematics model to compute joint angles for desired end-effector positions.
- ✅ Integrated trajectory planning for smooth motion execution.
- 🔄 Current Task: Implementing AI-based image recognition to detect welding patterns using a Raspberry Pi camera.
- 🔄 Upcoming: Controlling a welding nozzle mounted on the robotic arm to follow detected patterns.
Next Steps:
✔ Optimize real-time servo control for welding accuracy.
✔ Improve AI image recognition model for better detection.
✔ Test full automation loop from pattern detection → kinematics → welding execution.
📌 Final Thoughts & What’s Next
This project has been an exciting blend of CAD design, ROS2 simulation, real-world control, and AI automation. The next major goal is to finalize the AI-based welding system and conduct real-world testing.
💬 Stay tuned for updates! 🚀