Integration of Computer Vision and Iot in a Robotic Arm–Conveyor Belt System with Stability-Enhanced Servo Control for Object Manipulation

This study investigates the integration of computer vision and IoT technologies in a robotic arm and conveyor belt system to enhance efficiency and flexibility in object classification and manipulation. The system comprises three main modules: a robotic arm, a conveyor belt, and a computer vision unit. Each module performs a specific function and works in coordination to carry out the overall task, which includes detecting the target object on the conveyor belt and controlling the robotic arm to pick and place it onto a pallet. In contrast to conventional approaches that rely on complex models and constant wired communication, this study proposes a compact and resource-efficient solution by combining three key components: a lightweight object detection model, a servo-based motion control mechanism for smooth operation, and an intermittent wireless communication framework using the WebSocket protocol. To further reduce computational load, a non-continuous activation mechanism was applied to the vision module without compromising system responsiveness. Training with ±30% brightness augmentation further improved the object detection accuracy, contributing to a higher mean average precision (mAP). To identify the most suitable detection model for the proposed system, various architectures, including YOLOv8, YOLOv10, YOLO-World, YOLO11, and RT-DETR were evaluated. Among them, YOLO11n was identified as the most suitable, achieving 99.5% mAP with only 2.6 million parameters. These results demonstrate that the proposed system offers high accuracy while maintaining lightweight and flexible deployment, making it effective for real-time industrial applications.