FedBHAD: Energy-Efficient Federated Learning for Black Hole Attack Detection in RPL-Based Low-Power IoT Networks

The internet of things is a network of connected devices that share and send information over the internet, frequently in resource-constrained situations. These are often built using the routing protocol for low-power and lossy networks (RPL), face significant security problems because of their limited computing power, and have energy constraints. The objective of this study is to design an efficient and lightweight mechanism for detecting black hole attacks on RPL-based internet of things networks. The proposed framework presents a distributed collaborative learning framework to reduce the processing load on central nodes while enhancing real-time threat detection. The novelty of the present work lies in integrating distributed learning with feature-based anomaly detection tailored for RPL environments, thereby improving IoT network security while reducing communication and energy overhead. A customized data retrieval algorithm is developed with the Cooja simulator’s configuration and extracts essential network parameters, including rank, expected transmission count, power consumption, forward count, and reception count. The analysis of this dataset allows the detection of black hole attacks. The research analysis indicates that the proposed framework achieves 99.6% detection accuracy, surpassing existing machine learning and deep learning techniques and offering enhanced security, reduced overhead, and lower computational needs.