Federated Learning for Food Supply Chain Management

The escalating global population has amplified concerns regarding the provision of high-quality food. In response to this issue, various food supply chain monitoring systems have been developed. This paper introduces an innovative approach that combines blockchain technology with a multi-trust package-based model to enhance the reliability and efficiency of supply chain interactions. Our proposed framework includes a Low-Power Wide-Area Network (LPWAN) IoT setup beside methodologies to enhance sensor power consumption. It employs Hyperledger Fabric for the secure and immutable recording of data. Furthermore, the framework utilizes a federated multi-task learning approach for anomaly detection and the prediction of time-series sensory data, thereby fostering robust trust relationships among supply chain participants.