An RFID-augmented information retrieval technique for AI-enabled IoT devices ensuring Agri-Food Traceability and Anti-Counterfeiting

Globalized agri-food supply chains face rising risks of traceability gaps and product fraud, especially during early production stages where environmental conditions affect safety and quality. This work introduces a compact and low-cost, power efficient MCU-based Edge-AI system that enhances real-time traceability and anti-counterfeiting from harvest to consumer. Environmental data are collected during harvesting, and the information saved into the RFID tag is generated directly from these data through an embedded AI model, enabling local, connectivity-free decision-making. The system is designed to run efficiently on microcontrollers, making it fully deployable in field conditions. The resulting information is securely stored in the RFID tag, which acts as a portable, tamper-evident data carrier. A dual-scan mechanism and backend anchoring ensure authenticity, while a blockchain layer provides immutable record-keeping in a simulated environment. A prototype applied to grape harvesting demonstrates high model accuracy, fast on-device inference, low power usage, reliable RFID operations, and ease of integration with a web-based traceability platform. These results show that combining Edge-AI with RFID and blockchain provides a scalable and practical solution for improving transparency and protecting agri-food products against counterfeiting.