Secrecy and Covertness for UAV-Assisted Wireless Transmission With Non-Gaussian Noise

With the development of communication technology and the wide application of unmanned aerial vehicles (UAVs), the problem of covert and secure transmission in UAVS communication with ground is becoming more and more important. Aiming at the covert secure transmission problem of non-Gaussian noise-assisted UAVs networks under the ground node interference collaboration scenario, a joint trajectory and power optimisation scheme based on the dual-delay depth deterministic policy gradient technique is proposed to improve the throughput of covert secure transmission. Firstly, based on the system model of UAV line-of-sight(LoS) link, an objective function to maximise the average secure covert communication throughput from the start point to the end point is constructed under the constraints of covertness and security. Then, this optimisation problem is transformed into a single-intelligence Markov Decision Process (MDP), and a deep reinforcement learning algorithm is constructed to achieve the optimal decision, which leads to the optimal solution of the objective function. In this method, the problem of insufficient policy updates during training is effectively solved by incorporating a unique reward shaping mechanism into the MDP, which in turn effectively reduces the local convergence of the algorithm and achieves fast convergence of the training network. Simulation results show that the proposed method has faster convergence and higher rewards than the baseline scheme based on average power and without reward shaping, and achieves a higher average security concealment throughput in the UAV network scenario assisted by non-Gaussian noise at ground nodes.