Single and bulk identification of plastics in the recycling chain using Chipless RFID tags

Abstract

We examine the use of chipless RFID tags for plastic identification in two applications in the recycling chain. First, on a conveyor belt system, where tagged plastics pass by a reader one at a time. We demonstrate that our approach can successfully differentiate between 2 plastic types with over 90% accuracy and is agnostic to tag orientation or contaminants such as oil and water. In doing so, we show that the technology works just as well or is superior to comparable studies using optical methods. Second, for contents estimation, where we estimate the relative fraction of two types of plastics in a bulk mixture. We show that we are able to detect pure or homogeneous bales of plastic with over 90% accuracy and estimate the content of non-homogeneous bales with 65-75% accuracy. This could help recycling plant managers prioritize sorting operations based on composition of incoming shipments with more information than they have currently. Future directions of research are also discussed.