Project Abstract

This project on a chess-playing robot is driven by the growing interest in integrating artificial intelligence with robotics, responding to both technological advancements and the increasing accessibility of computer vision techniques. The project is motivated by the potential to revolutionize interactive gaming and automation in robotic tools, bringing innovation to traditional board games and modern technology into focus. The research proposition centres on developing a system that not only recognizes a physical chessboard and its pieces using computer vision and a trained detection model but also calculates optimal moves using the minimax algorithm, which merges perception with decision making in a physical application. The project uses a multi-layered approach that integrates image processing for board detection, machine learning for precise piece identification, and classical game theory algorithms to generate competitive chess moves, combined with the use of the Nyrio Ned 2 robotic arm to physically execute these moves. Key findings show that the system successfully identifies board configurations and chess pieces in real time, effectively computes strategic moves, and reliably manipulates the robotic arm to perform game actions. The project contributes to the field by demonstrating an effective implementation of computer vision, machine learning, and robotics in a real-world application, highlighting the practical potential of merging these fields and setting the stage for future enhancements in automated gameplay and interactive robotic systems.

Keywords: Robotics, Computer Vision Chess Detection, Chess strategy computation

 Conference Details

Session: A

Location: Sir Stanley Clarke Auditorium at 11:00 13:00

Markers: Daniele Cafolla, Deepak Sahoo

Course: BSc Software Engineering 3yr FT

Future Plans: I’m looking for work