Double Deep Q-Network-Based Smart Channel Hopping for Reactive Jamming Mitigation in IoT Networks: A Reliability and Energy-Aware Approach

Abstract

Reactive jamming attacks are prevalent in low power Internet of Things networks and are severely affecting the performance of IoT nodes. Existing anti-jamming studies focus primarily on communication reliability. Limited attention has been given to the joint optimization of packet delivery ratio (PDR), energy consumption, and network lifetime in mobile IoT networks under reactive jamming attacks. This study proposes a double deep Q-network (DDQN)-based smart channel hopping model (SCHM) that improves communication reliability while reducing energy expenditure and extending network lifetime. We use a Markov Decision Process (MDP) for modelling the interaction between the jammer and the legitimate IoT node. We develop a DDQN-based SCHM algorithm using the Python programming language. We implement the DDQN-based SCHM approach in the NS-3.41 environment. We evaluated the effectiveness of DDQN-based SCHM against Deep-Q-Network (DQN)-based hopping, non-hopping, and random hopping approaches. The results showed that our proposed DDQN-based SCHM approach achieved an average PDR of 99%, experienced the lowest energy consumption, and prolonged the network lifetime of IoT nodes. The proposed DDQN-based SCHM approach improves communication reliability and energy efficiency of IoT nodes.