Tic-Tac-Toe Robot : Automated Tic-Tac-Toe Game With Robotic Control

Abstract: This thesis covers the design, construction and programming of an automated tic-tac-toe game with robotic control and image processing. Research shows that robotic systems, over time, have transitioned from manufacturing tools to service tools with a large focus on human interaction, a central aspect- of this project. The research questions of the project are centered around the use of inverse kinematics and difficulties with using color recognition to identify player moves. Furthermore, a theoretical analysis of forward and inverse kinematics, image processing, the minimax algorithm and electrical components is presented. The prototype was developed using a Raspberry Pi, servo motors and a camera module to identify player moves. Multiple tests with regards to speed, accuracy and reliability were performed to evaluate the performance of the prototype. The accuracy tests showed minor positional deviations in the movement of the prototype. Additionally, the tests of the image processing algorithm showed that lighting is important to attain reliable color recognition. The project resulted in a fully functional prototype that enables a game of tic-tac-toe to be played against an automated opponent. Potential future work includes an improved design, more efficient methods to determine the optimal move and alternative image processing algorithms.