This paper presents the design and early deployment of a digitally enabled circular value chain for Expanded Polystyrene (EPS) in an outermost European island region, developed within the ISLOPOL project under CIRCULOOS.
The solution combines a low-cost AI vision system, based on RGB cameras and edge computing, with a FIWARE-based local data space to improve the identification, traceability and monitoring of EPS throughout the value chain. Tested in a real waste-sorting environment, the approach connects waste collection and sorting with local industrial reuse, helping turn a difficult-to-manage waste stream into a traceable secondary resource.
Early deployment results demonstrate the feasibility of using affordable and interoperable digital technologies to support circular economy models in geographically constrained regions. The next phase will focus on further AI model validation, dataset expansion and the assessment of operational, environmental and circularity impacts.