Joint Resource Allocation and UAV Trajectory Design for Data Collection in Air-Ground Integrated IoRT Sensors Network With Clustered NOMA

Abstract

Compared with the terrestrial network, the air-ground integrated network composed of unmanned aerial vehicles (UAVs) and high-altitude platforms (HAPs) has the advantages of large coverage, high capacity, and seamless connectivity, which can provide effective communication services for the Internet of Remote Things (IoRT) sensors. In this article, considering two transmission modes for two types of data with different delay requirements, and the limited battery capacity of UAV, we formulate a joint resource allocation and UAV trajectory design problem in clustered nonorthogonal multiple access (C-NOMA) air-ground integrated IoRT sensors network to maximize the data collection efficiency. For the formulated nonconvex problem, the deep deterministic policy gradient (DDPG) method can solve it. However, the DDPG method has the Q-value overestimation problem; in order to alleviate the problem, the twin-delayed DDPG (TD3) method with a double critic network is applied, and a TD3-based resource allocation algorithm is proposed to solve the primal problem. Simulation results verify that the proposed algorithm has better performance in terms of improving the data collection efficiency than other benchmark methods.