An Intelligent Recycling Sorting System Based on Yolov8 Image Recognition and a Mobile Robotic Arm

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

With the rapid growth of global urbanization and changing consumption patterns, the quantity and diversity of recyclable waste have increased significantly. Conventional waste sorting methods, which rely heavily on manual labor, often suffer from low operational efficiency, high labor costs, and inconsistent classification accuracy. To address these challenges, this study proposes an intelligent recycling sorting system based on YOLOv8 image recognition and a mobile robotic arm. The proposed system integrates deep learning, computer vision, Internet of Things (IoT) communication, and automated control technologies to improve the automation level and operational efficiency of recycling classification. The proposed system employs the YOLOv8 object detection model for recyclable waste recognition. A customized dataset was established and annotated using Roboflow, and the trained model was deployed for real-time object detection through OpenCV image acquisition. The overall system consists of a PC-based vision recognition platform, a Jetson Nano edge computing controller, an ESP32 embedded control module, an MQTT message communication mechanism, and a WLKATA robotic arm. When the ESP32 detects that a recyclable object has entered the sorting platform, it triggers the recognition process via MQTT. After object classification is completed on the PC, the recognition results are transmitted to the Jetson Nano, which controls the robotic arm to grasp the target object and perform automatic sorting. Meanwhile, the mobile platform transports the object to the designated recycling area, and the entire system is monitored in real time through a Node-RED dashboard. Experimental results demonstrate that the proposed system successfully performs real-time object recognition, autonomous grasping, and automatic recycling classification with stable system integration and reliable operational performance. The results confirm that integrating YOLOv8, Jetson Nano, ESP32, MQTT, and a mobile robotic arm provides an effective intelligent recycling framework capable of reducing manual sorting requirements while improving recycling efficiency. Furthermore, the proposed system offers a practical reference architecture for intelligent recycling, smart manufacturing, and AIoT applications, and demonstrates strong potential for future deployment in smart environmental protection and circular economy applications.