A Federated Learning-Based Intrusion Detection Framework for Blockchain-Enabled IoT Supply Chains

The newest Internet of Things (IoT) networks in supply chain operations not only provide massive amounts of realtime information but also present new cyber-physical security challenges. This study presents a multidisciplinary simulation framework that merges blockchain and machine learning (ML) to boost the security and efficiency of blockchainbased IoT supply chain networks. In this regard, we consider federated learning (FL)-trained deep neural networks for distributed anomaly detection on IoT device data collections and report anomalies on a permissioned blockchain database to achieve tamper-proof auditability measures. When we simulated the network (using NS3 and a blockchain simulator) under various trouble scenarios, the ML-enhanced system could detect cyber-attacks (e.g., data exfiltration, DDoS, and malicious device behavior) with $\mathbf{{\gt} 9 5 \%}$ accuracy and also considerably reduced the false alarms compared with the baseline methods. Blockchain only provides data visibility and traceability for supply chain events. The following are our contributions: (1) proposing a new federated ML framework for distributed anomaly detection in IoT supply chains, (2) designing a layer on the blockchain to store alerts and trained models, and (3) performing a comprehensive simulation study of the overall system, which demonstrates how combining both systems surpasses non-ML or non-blockchain configurations. The results indicate that a robust collaborative ML-based IDS improves incident response, and the blockchain guarantees the trust and auditability of IoT Data streams in supply chains. This study lays the groundwork for the future realworld deployment of secure, smart IoT supply chain networks.