Better data basis, flexible interfaces and new sorting criteria
The project partners want to pave the way for a sensor-based sorting technology for a broad field of application in plastics recycling by upgrading, integrating and testing the developed multi-sensor system in a sorting system. For this purpose, novel sensor data or information fusion methods have to be developed, especially by machine learning methods. The new process, which is linked to the digital twin, is to be economical, enable high degrees of purity for plastics sorting and, in addition, provide robust material characterization including black plastics and previously unrecorded criteria (e.g. aging condition).