Resource Allocation and Robust Trajectory Design for Dual UAV-Assisted Secure Communications

This paper introduces a dual unmanned aerial vehicle (UAV)-assisted communication system to facilitate secure information transfer between an Internet of Things (IoT) device and a base station in the presence of multiple eavesdroppers. We consider a more realistic scenario where the legitimate nodes have imperfect knowledge about the eavesdroppers' locations that can intercept information through air-to-ground and ground-to-air links. Both UAVs and energy-constrained IoT device harvest energy from radio-frequency signals transmitted by a nearby power beacon. Utilizing the harvested energy, one UAV acts as an aerial relay while the other UAV broadcasts jamming signals to ensure information secrecy. In the considered system, we formulate an average secrecy rate maximization problem by jointly optimizing the time-splitting ratio, transmit powers of IoT device and UAVs, and trajectories of both UAVs. Moreover, we account for practical constraints such as collision avoidance, limited battery capacity, energy harvesting and information causality, and UAVs' maximum velocity and flying radius. The formulated problem is non-smooth and non-convex, which we first transform into a smooth objective function without affecting its optimality. Then, we relax the eavesdropper's uncertainty constraint by using Jensen's inequality and obtain upper bound of the wiretapping rate. Further, we decompose the original optimization problem into five subproblems using a block coordinate descent algorithm and transform them into convex form via successive convex approximation. These subproblems are then solved using the convex optimization tool CVX, simplex, and subgradient descent methods. Lastly, we show the effectiveness of our framework by comparing the secrecy performance with benchmark schemes in simulation results.