Early Adaptive Intrusion Detection System for Mobile IoT Networks

In recent years, the security of the Internet of Things (IoT) has attracted significant attention from researchers due to its unique characteristics, such as device heterogeneity, resource constraints, and novel attack types specific to IoT environments. Furthermore, the presence of mobile attackers leads to another layer of complexity in IoT security. In such challenging network conditions, Intrusion Detection Systems (IDSs) have become increasingly vital. Traditional detection systems typically rely on assumptions of static network topologies and fixed attacker scenarios, often neglecting mobility and dynamic traffic patterns. To address this, the present study proposes an innovative IDS enhanced with an Evolutionary Dynamic Optimization (EDO) strategy, specifically designed to be effective in dynamic environments. By integrating Genetic Programming (GP) with the EDO mechanism, the proposed system evolves a detection algorithm capable of adapting to temporal changes in adversarial behavior. Moreover, the system incorporates Diversity Introducing (DI) and Transfer Learning (TL). The proposed IDS supports the rapid convergence of GP, enabling early adaptation to environmental changes. This study offers a significant contribution to the literature as the first approach to provide automated and early adaptation to environmental dynamics for detecting mobile attackers in RPL-based IoT networks.