ABSTRACT This paper presents a fully automated chess‐playing system that integrates computer vision, artificial intelligence, and robotics to enable autonomous gameplay. The system comprises three core components: (i) YOLOv8, a fine‐tuned deep learning model for chess piece recognition; (ii) Stockfish, a high‐performance chess engine for strategic move selection; and (iii) the Quanser QArm, a robotic manipulator for precise move execution. The methodology involves fine‐tuning YOLOv8 on a custom dataset, using FEN notation to represent board states, and executing moves via pre‐calibrated robotic waypoints. The system was evaluated in real‐world settings, achieving high detection accuracy and reliable robotic control, with the QArm completing 87% of moves on the first attempt and 97% by the second. Our system contributes a unified, modular architecture that enables reliable autonomous chess gameplay under real‐world conditions.
Alkhatib et al. (Wed,) studied this question.