Scrabot: a companion scrabble player robot

Abstract

Due to the increasing advantages of playing board games on mental and physiological health, especially for people with brain diseases, the need for a companion robot increased. Hence, ScrabBot, a Whole Arm Manipulator (WAM) robotic arm from Barrett technologies is currently under development to be the perfect companion and game opponent. This robot is 7- DOF and consists of three main parts: Barrett WAM arm to physically move tiles and execute the gameplay, a pyKinect Azure camera for computer vision purposes, and an AI agent that would serve as the brain of the robot. For the robotic arm, a controller is deployed on Robotic Operating System (ROS) using an external IK solver. For computer vision, OpenCV is used for letter detection and an OCR library is to be used for character recognition. For the AI agent, our project will develop an innovative AI agent using novel techniques such as reinforcement learning algorithms to improve on currently-existing approaches in the literature. Moreover, our AI agent’s design will push the robot towards playing more frequently-used words in order to make the robot sound more like a human companion. This robot will not be limited to playing Scrabble but will be able to play any similar game like chess. However, for the purposes of the Capstone Project delivery, we will only be developing an AI agent for Scrabble.