RecycLight: Enhancing PET Bottle Collection for Sustainable Waste Management in Universities with the Use of Optical Recognition in Reverse Vending Machine

Polyethylene terephthalate (PET) bottle waste remains a pressing environmental challenge, particularly in the Philippines, where insufficient recycling infrastructure and low public awareness contribute to widespread pollution. This study examines the deployment of artificial intelligence (AI)-driven reverse vending machines (RVMs) equipped with optical recognition technology as an innovative approach to enhance PET bottle collection and recycling in universities. Conducted in the National Capital Region (NCR), the research adopts a mixed-methods approach, integrating surveys and experimental testing of an RVM prototype to evaluate its performance and impact. The survey, with a sample size of 408 respondents (77.5 % students, 19.5 % faculty, and 1.5 % staff), revealed high engagement among university stakeholders, with 76.4 % aged <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$18-24$</tex> years. Key findings from multiple regression analysis identified machine learning algorithms (<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\beta=0. 0 9 8, p=0. 0 2 2$</tex>) and reliable power supply (<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\beta=0.108, p=0.008$</tex>) as significant factors influencing system performance. Experimental testing demonstrated a recognition accuracy rate of 92 % for PET bottles without labels or lids, with the system processing three bottles per user efficiently during trials. Furthermore, paired ttest results highlighted significant improvements in collection speed and accuracy (<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$p &lt; 0. 0 5$</tex>), validating the system's potential to streamline recycling operations. The study underscores the effectiveness of AI-driven RVMs in addressing critical challenges in PET bottle waste management, promoting recycling behaviors through reward systems, and aligning with the United Nations' Sustainable Development Goals (SDG 12: Responsible Consumption and Production). Practical recommendations are provided for integrating such systems into university recycling frameworks, emphasizing their scalability and environmental benefits.