RIS-Assisted Air-to-Ground Communications with Non-Orthogonal Multiple Access

This paper investigates the reconfigurable intelligent surface (RIS) aided air-to-ground uplink cellular network with non-orthogonal multiple access (NOMA). A major novelty of the proposed framework is that it introduces the new paradigm of flexibility on efficient spectrum sharing between aerial and ground cellular users by taking advantage of the aerial users’ high mobility, reconfigurable wireless environment as well as power-domain multi-user access. With the aim of maximizing network sum rate while guaranteeing users’ quality-of-service (QoS) requirements, a joint problem of unmanned aerial vehicle (UAV) trajectory and RIS configuration design is formulated. The formulated problem is a sequential decision making one across multiple coherent time slots. Moreover, considering ground users’ random movement in practice, the decision making process involves uncertainties. To solve the formulated problem efficiently, the deep reinforcement learning algorithm is adopted to intelligently adjust the UAV’s trajectory and RIS configuration simultaneously. Numerical results demonstrate that the network sum rate is significantly improved with the proposed NOMARIS scheme compared to conventional orthogonal multiple access (OMA) communication and the case where no RIS is deployed.