Energy-Efficient Optimization Scheme for RIS-Assisted Communication Underlaying UAV with NOMA

Abstract

Unmanned aerial vehicles (UAVs) and reconfigurable intelligent surface (RIS) are the emerging technologies for 5G and beyond networks. These two techniques reduce inter-user interference and enhance the network's coverage performance. Despite this advantage, these two techniques are not able to satisfy the diversified quality of service (QoS) requirements of cellular mobile users under the presence of existing multiple access schemes. To tackle this issue, we integrate non-orthogonal multiple access (NOMA) with both these techniques. In this paper, our goal is to maximise the energy efficiency (EE) of the overall network by optimising the powers of UAVs and the phase shift matrix of RIS. The formulated problem is in a mixed-integer non-convex programming form. So, to solve this problem, a deep deterministic policy gradient (DDPG) approach is used in a centralised manner under a time-varying channel. The proposed NOMA-RIS scheme for multi-UAV networks achieves higher EE than the orthogonal multiple access (OMA)-RIS and random selection schemes, according to numerical results.