Design and Development of a 6 DOF Arbiter Robotic Arm for Games of the General (Salpakan) Using Yolov8

This study presents the design and development of an autonomous 6 degrees of freedom (DOF) arbiter robotic arm for the Filipino board game “Game of the Generals” or “Salpakan.” The robotic arm is equipped with a parallel jaw gripper and made intelligent using deep learning models from the You Only Look Once (YOLO) family, specifically YOLOv8, YOLOv9, and YOLOv10. These models enable real-time object detection and decision-making, allowing the arm to enforce game rules, determine the outcome of challenges, and retrieve losing pieces. A dataset consisting of 30 classes was created, representing the various pieces used by opposing players. The models were trained using a batch size of32 and 100 epochs. The performance of the YOLO models was evaluated using confusion matrices, with results showing that YOLOv8 achieved the highest accuracy, making it the most suitable model for this application. YOLOv8 produced the highest mean average precision (mAP) scores, with mAP50 and mAP50-95 values of 99.5% and 85.4%, respectively. This integration of AI and robotics enhances the fairness and functionality of strategic board games, providing an automated and impartial arbiter for the “Game of the Generals.” This research demonstrates the potential of combining deep learning and robotics to innovate traditional board games, ensuring accuracy and efficiency in game management.