A.R.I.S.: Automated Recycling Identification System for E-Waste Classification Using Deep Learning
Dhruv Talwar, Harsh Desai, Wendong Yin, Goutam Mohanty, Rafael Reveles
TL;DR
The paper tackles inefficient e-waste recycling by integrating a real-time, deep-learning based classifier with a low-cost, inline sorting mechanism. It presents A.R.I.S., a multi-camera YOLOx-based system that classifies shredded e-waste into metals, circuit boards, and plastics and drives a pneumatic paddle sorter via a PLC-OPC-UA control loop. The approach achieves an overall mAP of $82.2\%$ on a held-out test set, with high recall for metals ($86.3\%$) and circuit boards ($94.1\%$) and high precision for plastics ($99.7\%$), translating to practical recovery improvements ($89\%$ for metals, $85\%$ for boards, $79\%$ for plastics) at about $5\,\mathrm{kg/s}$ throughput. The work demonstrates a scalable, cost-effective pathway toward industrial-grade automated e-waste sorting, with potential for broader deployment and further gains by addressing plastics and small-fragment detection.
Abstract
Traditional electronic recycling processes suffer from significant resource loss due to inadequate material separation and identification capabilities, limiting material recovery. We present A.R.I.S. (Automated Recycling Identification System), a low-cost, portable sorter for shredded e-waste that addresses this efficiency gap. The system employs a YOLOx model to classify metals, plastics, and circuit boards in real time, achieving low inference latency with high detection accuracy. Experimental evaluation yielded 90% overall precision, 82.2% mean average precision (mAP), and 84% sortation purity. By integrating deep learning with established sorting methods, A.R.I.S. enhances material recovery efficiency and lowers barriers to advanced recycling adoption. This work complements broader initiatives in extending product life cycles, supporting trade-in and recycling programs, and reducing environmental impact across the supply chain.