Within the framework of the new European directives on the circular economy, the issue of plastic recycling, particularly PET bottles, remains one of the major environmental challenges. One of the main obstacles to recycling is the need to return bottles intact, a requirement that complicates domestic waste management.
For this reason, researchers Rosca and Stancu have devised an innovative solution that combines technology, simplicity and rewards for virtuous citizens. The proposal is the PRS (PET Recycling System), an intelligent, low-cost system that allows crushed bottles to be collected without compromising the efficiency of separate waste collection. Based on an AIoT (Artificial Intelligence of Things) platform, the prototype uses a combination of artificial vision, weight sensors and decision-making algorithms. At the heart of the process is the PBIA (PET Bottle Identification Algorithm), which is capable of recognising bottles even if they are deformed, accurately distinguishing them from other similar materials. To ensure reliability, each object is only accepted if it exceeds a “confidence” threshold (PIC), calculated by cross-referencing the visual result with the weight result.
A key element is the active involvement of citizens. Thanks to the ACTLR (Algorithm for Citizens’ Trust Level by Recycling), each user has their own profile and trust score, which increases with correct behaviour and decreases in the event of errors or attempts to dispose of incorrect materials. In the event of abuse, the system also takes a photo of the user, protecting the integrity of the process. The score, calculated according to an exponential formula, rewards those who recycle responsibly and penalises incorrect behaviour, even if sporadic. Through a clear and accessible software interface, citizens can check their points, receive notifications on the status of the container and access rewards made available by affiliated shops.
The system, designed for everyday use, has already been tested with interesting results: up to 96% accuracy in visual recognition, reaching 98% thanks to the PBIA algorithm. All this at a low cost and with a structure that can be easily replicated on an urban scale.
In summary, Rosca and Stancu’s study proposes a concrete, intelligent and participatory way to simplify domestic PET collection. Less clutter in the home, more rewards for citizens and a system ready to bring real benefits to the environment and the community.
Bibliographical references: Rosca, C.-M.; Stancu, A. Innovative AIoT Solutions for PET Waste Collection in the Circular Economy Towards a Sustainable Future. Appl. Sci. 2025, 15, 7353. https://doi.org/10.3390/app15137353
