Robotic Hand

Robotic hand is one of my favorite projects because it contains a variety of solutions. My aim was to create a robot which can play rock, paper, scissors with a human. Presented version of my project was built using an Arduino Mega board, 3 Micro Servos, IR Receiver and hand rig. Another part of this project is camera which identifies hand position. The camera is connected to a Neural Network which identifies if the hand represents a rock, paper or scissors. The neural network was trained on 100s of images of hand gestures that depict rock, paper and scissors which had to be processed in different lighting conditions and calibrated to make the model more efficient. Finally, the third important part is the algorithm. My algorithm is based on the neural networks prediction. The code is processing result and converts it to a format which Arduino can understand. In the last step the algorithm sends result to Arduino which controls robotic hand and gives instruction to take the relevant responding position.