Welcome to the LEAP Hand ROS1 SDK - Rock, Paper, Scissor

demo.mp4

Prerequisites

• Operating System: Ubuntu 20.04
• ROS Version: Noetic

Install Dependencies

pip install empy==3.3.4 catkin_pkg pyyaml rospkg
pip install dynamixel_sdk numpy

Creating a ROS Workspace

1. Install the dependencies and navigate to your ROS workspace.
2. Create a new workspace:

   cd ~/leap_hand_ws/src

3. Clone Leap-Hand-Robotics inside the ~/leap_hand_ws/src directory

   git clone https://github.com/Demolus13/Leap-Hand-Robotics.git leap_hand --recursive

4. Create Executable Files inside the ~/leap_hand_ws/src/leap_hand directory

   chmod +x leaphand_node.py
   chmod +x rock_paper_scissors.py

Building the workspace

1. Build your workspace:

   cd ~/leap_hand_ws
   catkin_make

2. Source your workspace:

   source devel/setup.bash

Connect to the Leap Hand Hardware

To Connect

• Connect 5v power to the hand (the dynamixels should light up during boot up.)
• Connect the Micro USB cable to the hand (Do not use too many USB extensions)
• Find the USB port using Dynamixel Wizard

To Launch

1. Terminal 1

   roslaunch leap_hand example.launch

2. Terminal 2

   rosrun leap_hand rock_paper_scissors.py

Finally Enjoy Playing the Game !!!

Note: We have already saved the gestures of rock, paper, and scissors in the rock_paper_scissors.py

# Define the joint positions for rock, paper, and scissors
self.states = {
    "rock": np.array([3.1416, 4.1888, 4.5553, 4.4157, 3.1416, 4.1190, 5.1487, 4.2412, 3.1416, 4.2237, 4.7124, 4.4506, 2.6005, 1.5184, 4.6775, 4.4157]),
    "paper": np.array([3.1416, 3.1416, 3.1416, 3.1416, 3.1416, 3.1416, 3.1416, 3.1416, 3.1416, 3.1416, 3.1416, 3.1416, 3.1416, 3.1416, 3.1416, 3.1416]),
    "scissors": np.array([3.1416, 3.1416, 3.1416, 3.1416, 3.1416, 3.1416, 3.1416, 3.1416, 3.1416, 4.2237, 4.7124, 4.4506, 2.6005, 1.5184, 4.6775, 4.4157])
}

Additional Steps: To train your own hand gesture model

1. Capture Images for Training using the script capture_images.py
   • On executing the python file the webcam will open can you can save your hand gesture by pressing Enter-Key to capture.
   • Captured images will be stored in the data/raw
2. Execture preprocessing using the script preprocess_data.py
   • The preprocessed .csv files will be created in data/processed
3. Train the Model using the script train_model.py
   • The trained model and the label encodings will be stored in models
4. Evaluating model using the script evaluate_model.py
   • The webcam will open and you can perform different stored gestures to check the accuracy of the model

Follow the above steps for simulating on the Leap Hand Hardware