<![CDATA[This research designs and builds a wheeled soccer robot using YOLOv8 for real-time ball detection and distance estimation, aiming to improve efficiency in technology competitions. The system includes Arduino Uno R3, Raspberry Pi 3 model b, detection system, and navigation design. 691 ball image use as dataset that consist of 552 image as training dataset and 249 image as valid dataset. YOLOv8 demonstrated exceptional reliability in ball detection during testing, achieving an average accuracy of 100%, 100% precision, and 94% recall. Navigation testing toward the ball had an acceptable average error of 8.0466%. The results confirm that YOLOv8 is excellent for simplifying high-accuracy ball detection and distance estimation in wheeled soccer robots. Future work should consider a higher-spec Raspberry Pi, a high-resolution camera, additional sensors, and advanced systems to improve detection and obstacle avoidance (opponent robots, goal).]]>
Copyrights © 2025
