(Igor Sikorsky Kyiv Polytechnic Institute, 2024) Junwei Ren; Boichenko, Sergii
Plastic recycling plays a pivotal role in advancing a circular economy. Effective mechanical recycling requires plastic waste to first be sorted by type, color, and dimensions to recover high-quality material for further processing, while chemical recycling requires excluding certain plastics that can impair the process. Sorting at Materials Recovery Facilities increasingly relies on advanced automated technologies, with accurate identification of plastic types typically achieved through spectroscopic analysis enhanced by machine learning and artificial intelligence. This paper outlines recent advances in integrating machine learning and AI into plastic recycling, reviewing current identification and sorting technologies, and drawing on publicly available data to assess commercial sorting equipment in terms of sensors used, sorting efficiency across plastic types, main applications, capacity, and precision. Despite significant progress, challenges remain in separating plastic films, dark plastics, and multi-polymer materials, and the field continues to evolve with new developments expected to further enhance sorting automation.