Optimizing Downlink Communication in a Multi-STAR-RIS-Assisted Multi-Antenna AAV Network

Abstract

Autonomous aerial vehicles (AAVs) are crucial for enhancing global connectivity, but the AAV transmissions are vulnerable to limited onboard energy and signal blockages in dense urban areas. Reconfigurable intelligent surfaces (RISs) can mitigate blockages, while allowing for fewer on-board antennas and reduced energy consumption for AAVs. In this letter, our objective is to study whether the distributed RIS network can achieve the gains comparable to a multi-antenna AAV, thus enabling reduced on-board energy consumption. To this end, we develop a framework to optimize multi-user scheduling, amplitude and phase shifts of simultaneous transmission and reflection (STAR)-RISs, and AAV beamforming in a distributed STAR-RIS-assisted multi-antenna AAV network. In this context, the sum-rate maximization problem is non-convex due to the interdependence among beamforming, phase shifts, and scheduling variables. To solve the problem, we decompose it into three sub-problems. We use semi-definite programming and integer constraint relaxation to solve the phase shifts and scheduling optimization, and apply standard successive convex approximation for beamforming optimization. We then employ alternating optimization to iterate until convergence is achieved. Our findings offer insights into scenarios where a distributed STAR-RIS network can achieve performance gains comparable to a multi-antenna AAV network.