Authors: Piyali Ghosh , Dr. Dhirendra Kumar Tripathi
Abstract: With the accelerated growth of the Internet of Things (IoT), providing secure, scalable, and privacy-preserving communication has become a serious issue. Current routing protocols such as AODV, DSR, and HSRP are incapable of addressing the complex needs of today's IoT systems, particularly in large-scale, heterogeneous, and energy-constrained systems. This paper introduces the Federated Hybrid Secure Routing Protocol (F-HSRP)—a new paradigm that combines federated learning, trust-based routing, AES-256 encryption, and blockchain-aided route verification to address these issues holistically. F-HSRP utilizes a light-weight Convolutional Neural Network (CNN) at the edge of IoT networks. With federated learning, local anomaly detection models are trained at the nodes, maintaining data privacy and facilitating real-time accurate threat detection. Routing decisions are informed through a composite trust score based on node behavior, residual energy, and anomaly scores. Secure data transfer is enabled by AES-256 encryption and a lightweight Proof-of-Authority (PoA) blockchain process that guarantees tamper-proof route verification without imposing substantial overhead. The protocol is tested with the Bot-IoT dataset and a hybrid simulation platform that integrates NS-3 and TensorFlow Federated. The results indicate F-HSRP outperforms conventional protocols with 96.3% anomaly detection rate, 27% energy efficiency improvement, and better packet delivery and delay metrics. It also successfully fights blackhole, replay, and Sybil attacks. By integrating federated intelligence, cryptographic security, and blockchain consensus, F-HSRP offers a strong, energy-efficient, and privacy-enhanced routing solution for real-time IoT applications in smart cities, industrial control, healthcare monitoring, and military systems.
