A DRL-based Scheme for Denial-of-Charge Attacks in Wireless Rechargeable Sensor Network

With the emergence of wireless energy transfer technology, Wireless Rechargeable Sensor Network (WRSN) plays an important role in target sensing and objective monitoring. Most research work focuses on charger scheduling algorithms, while the issue of charging security is received limited attention. This paper develops a novel Denial-of-Charge (DoC) attack, aiming to provide an adversary model for network security. We study how to schedule a malicious Mobile Charger (MC) to disrupt network functionality while avoiding detection by the Base Station (BS). A Deep Reinforcement Learning (DRL) algorithm is developed to maximize the loss rate of monitored Points of Interest (POIs), called Denied of Charge through Dueling Double DQN (DoC-D3). Finally, we evaluate the proposed DRL algorithm through several simulation experiments, and the results demonstrate a significant increase in target loss compared to other algorithms.