Rock paper scissors robot arm

Gage Grant; Jason Katz

doi:10.64804/rt404f46

Authors

Gage Grant Science & Engineering Magnet Program , Manalapan High School
Jason Donovan Katz Science & Engineering Magnet Program , Manalapan High School

DOI:

https://doi.org/10.64804/rt404f46

Keywords:

rock, paper, scissors, robot hand, OpenCV, machine vision, cues, machine learning, Raspberry Pi, Arduino, 3D printing, multi-armed bandit problem

Abstract

Rock Paper Scissors is one of the most recognizable, easy, and popular hand games out there. And while it is sometimes fun to play as a last resort in an incredibly boring situation, winning roughly 50% of the time isn’t good enough for us. We want to not only win, but crush the competition. Humans are terrible at being random, and we want to make a robotic hand to exploit this human weakness.

Over the course of this semester, we have created a robotic hand that will beat humans at rock paper scissors. By using OpenCV, the Multi Armed Bandit Problem, and some open sourcing from rpscontest.com, we were able to create a program that can recognize the human’s move and calculate the best strategy against it. And with 3D printing, some servos and basic electronic parts, we created a basic robotic hand that utilizes a Raspberry Pi and Arduino to make the move to crush the human, in the game.