Welcome to the LEAP Hand SDK

👉 More info: LEAP Hand Website

---

Software Setup

• See these folders for setup details:
  • Python API
  • ROS API
  • ROS2 API
  • Useful Tools

---

🔌 Hardware Setup

• Connect 5 V power to the hand (Dynamixels should light up on boot).
• Connect the Micro‑USB cable (avoid multiple USB extensions).
• Use Dynamixel Wizard to find the correct port.
  ➡️ Put that port into main.py or ros_example.py.
  ⚠️ You cannot have Dynamixel Wizard open while using the API (the port will be busy).
• On Ubuntu, find the hand by ID at /dev/serial/by-id (persistent across reboots).
• sudo chmod 666 /dev/serial/by-id/(your_id) to give serial port permissions.
• Official support: Python and C++ and ROS/ROS2. Other languages can use the Dynamixel SDK.
• To improve latency on Ubuntu:
  • Adjust USB Latency Settings
  • Tune the Dynamixel Python SDK
  • Set Return Delay Time (Control Table Register 9) from 250 µs to 0 µs.
• If you are using the full hand, you can raise the current limit from 300 mA to 550 mA in the API for increased strength!

---

🤖 Functionality

• Leap Node allows commanding joint angles in different scalings.
• You can read position, velocity, and current.
• Query limits:
  • Position only: ≤ 500 Hz
  • Position + velocity + current: ≤ 500 Hz
    (Higher rates can slow USB communication.)
• Default control: PID (up to current limit).
  Velocity and current control also supported—see the motor manual.
• Current limits:
  • Lite: ≈ 300 mA
  • Full: up to ≈ 550 mA
  • By default the API is at 300, you can raise it to 550mA on the Full hand!!!
• Jittery hand? ➡️ Lower P/D values.
  Weak hand? ➡️ Raise P/D values.

---

🛠️ Troubleshooting

• Motor off by 90°/180°/270° → Remount the horn.
• No motors show up → Check serial port permissions.
• Some motors missing → Verify IDs and U2D2 connections.
• Overload error (motors flashing red) → Power cycle. If frequent, lower current limits.
• Jittery motors → Lower P/D values.
• Inaccurate motors → Raise P/D values.

---

🔧 Useful Tools

• MANO → LEAP joint angle mapping.
• Bimanual Dexterity for Complex Tasks shows how to use Manus gloves with LEAP Hand.
• Have a useful tool to share? Pull requests welcome!
  (Or ask and I can add tools for you.)

---

Support

• Questions/issues: kshaw2@andrew.cmu.edu
• License:
  • Code: MIT License
  • CAD: CC BY‑NC‑SA (non‑commercial use with attribution)
• Provided as‑is, without warranty.

If you use LEAP Hand in research, please cite:

@article{shaw2023leaphand,
  title={LEAP Hand: Low-Cost, Efficient, and Anthropomorphic Hand for Robot Learning},
  author={Shaw, Kenneth and Agarwal, Ananye and Pathak, Deepak},
  journal={Robotics: Science and Systems (RSS)},
  year={2023}
}