A partnership involving nine brand owners, an independent test and research center, and two universities is joining forces for a two-year initiative. Their shared objective is to create and evaluate an advanced artificial intelligence (AI) decision-making system. This system aims to enhance the accuracy of sorting packages, particularly those currently misclassified, including differentiation between food and non-food packaging items. Embracing open innovation, the consortium is focused on preventing technological entrenchment. Upon completion, the resulting AI decision model will be readily available for widespread adoption in sorting facilities throughout Europe.
Perfect Sorting
Currently, Europe faces a pressing challenge with over 79 million tons of packaging waste generated in 2019, with only a 65% recycling rate. Plastic packaging waste alone accounts for more than 15 million tons, but only 41% of it is recycled. Traditional sorting methods, relying on limited parameters such as material type, rigidity, and color, struggle to efficiently separate packages, especially food vs. non-food or flexible multimaterial packaging. The Perfect Sorting project, initiated by a consortium comprising National Test Centre Circular Plastics, leading brands such as Danone, Colgate-Palmolive, Ferrero, LVMH Recherche, Mars, Michelin, Nestlé, PepsiCo, Procter & Gamble, and esteemed universities including Ghent University and Radboud University, aims to revolutionize packaging waste sorting using artificial intelligence (AI). By harnessing the expertise of nine brand owners, an independent research center, and two universities, the project focuses on enhancing the existing waste sorting practices in Europe.
Waste sorting, recycling & programming
The Perfect Sorting project stands out for its unique composition, bringing together diverse expertise in ecodesign, waste sorting, recycling, and AI programming. The National Test Centre Circular Plastics will rigorously test various packaging products using its advanced sorting line. The results will be utilized to refine the AI-based sorting model developed by Ghent University. This collaborative effort is not confined to the consortium; it also aims to engage technology providers and waste management companies to enhance the project's impact. The potential of AI in waste sorting lies in its ability to detect and sort products based on multiple attributes, such as color, application, shape, or material. This nuanced approach could significantly increase sorting efficiency and improve the quality of recyclates. For instance, AI could facilitate the precise separation of food and non-food packaging, ensuring that recyclates meet the necessary standards for reuse in food-grade applications.
AI decision model
At the project's culmination in two years, the partners aspire to implement the AI decision model successfully in an industrial sorting plant. Importantly, they seek an implementation method that is cost-effective and straightforward, enabling swift integration into existing sorting technologies. The consortium's ultimate vision is to democratize this AI decision model, making it widely accessible for sorting plants across Europe, thereby revolutionizing the continent's approach to packaging waste management.