A CNN Based Fast Picking Method for WEEE Recycling 2022 Procedia CIRP

With the growing global waste of electrical and electronic equipment (WEEE), efficient recycling methods are urgently needed to mitigate environmental and health concerns. This study aims to enhance the WEEE recycling process through a CNN-based grasp planning method specifically designed for random picking of unknown objects. Utilizing deep learning techniques, the proposed method addresses the limitations of current manual and semi-automatic recycling processes, which are inefficient and hazardous. The methodology incorporates convolutional neural networks trained on diverse datasets to improve grasping accuracy and operational efficiency in sorting and recycling WEEE. Results demonstrate a significant enhancement in the grasping performance compared to traditional techniques, achieving higher sorting accuracy and a reduced risk to workers involved in the recycling process. The findings also indicate that the CNN-based approach can be effectively integrated into existing automated recycling facilities, paving the way for a more sustainable and secure WEEE processing environment. Future research directions include optimizing the CNN architecture and expanding its applicability to various categories of e-waste. This work serves as a step towards the development of intelligent automated systems in WEEE recycling.