Key points are not available for this paper at this time.
To maintain a green society, efficient waste management is crucial. Traditional manual trash sorting presents several challenges, including inaccuracies in classification and potential health risks for workers. To address these issues, this paper proposes an intelligent and automated waste classification system that integrates deep learning with robotic kinematic control. Our approach significantly improves classification accuracy, speed, and reliability compared to manual sorting. A diverse dataset containing various waste objects, including durian peels, was collected and labelled by experts. Using deep learning, the system was trained to recognize and classify objects with high precision. A camera mounted on the end-effector of robot identifies the position and orientation of object, enabling the robot to precisely pick up and sort waste items. The key advancements of our approach include (i) development of a robotic waste classification platform that enhances sorting efficiency and reduces human involvement, (ii) implementation of a model-based learning approach that achieves rapid and accurate object detection, (iii) validation through real-world experiments, demonstrating the feasibility and effectiveness of the system in complex environments. Experimental results confirm that the proposed system significantly enhances waste classification accuracy and efficiency, paving the way for safer and more intelligent waste management in smart manufacturing and environmental sustainability applications.
Le et al. (Fri,) studied this question.