Dual UAV-Assisted Secure IoT Networks: Resource Allocation and 3D Trajectory Design

This paper investigates a dual unmanned aerial vehicle (UAV)-assisted communication system for enabling secure information transfer between a base station (BS) and an Internet of Things device against randomly located terrestrial eavesdroppers, modeled as a homogeneous Poisson point process. In this setup, both the UAV relay and the jammer harvest energy from the radio frequency signals transmitted by the BS and utilize it for relaying and transmitting jamming signals. We formulate the average secrecy rate maximization problem by considering the time-splitting ratio, transmit powers, and the three-dimensional trajectories of UAVs. Moreover, we account for practical constraints such as energy harvesting and information causality, UAVs' collision avoidance, limited battery capacity, etc. Considering the system dynamics, we reformulate the problem as a constrained Markov decision process and solve it with a soft actor-critic (SAC) algorithm. Specifically, the SAC agent learns an optimal policy by observing time-varying channel gains and spatial randomness of homogeneous Poisson point process eavesdroppers. Lastly, we demonstrate that our framework significantly enhances ASR performance compared to benchmark schemes through numerical analysis.