Deep reinforcement learning optimizes multi-UAV communication with intelligent surfaces
Deep Reinforcement Learning for BD-RIS-Assisted Multi-UAV RSMA Communications
Information TheoryEmerging Technologies
Summary
Communicating with multiple drones efficiently is hard because their wireless signals change quickly and power is limited. The authors study a setup using smart surfaces that can control signal reflection in advanced ways, helping drones communicate better. They create a method that plans drone paths, signal sharing, and how these surfaces reflect signals to maximize total data rates. Their deep learning approach works better than several existing algorithms, mainly by improving how messages are split and shared, with some added benefit from the smart surfaces.
What this means in practice
- •For drone network engineers: Plan drone flight paths and communication strategies to boost network throughput using advanced signal reflection surfaces and reinforcement learning.
- •For wireless system designers: Design multi-user communication systems incorporating smart reflecting surfaces and message splitting, optimized via deep reinforcement learning for aerial networks.
Authors
Aamer Mohamed Huroon, Ali A. Nasir, Daniel B. da Costa
Abstract
Beyond-diagonal reconfigurable intelligent surfaces (BD-RISs) extend conventional diagonal RISs by allowing controllable coupling among reflecting elements. This flexibility is attractive for aerial networks, where unmanned aerial vehicles (UAVs) face rapidly varying channels and strict power limits. In this paper, we study a hybrid BD-RIS-assisted multi-UAV rate-splitting multiple access (RSMA) downlink, in which UAVs within favourable BD-RIS coverage use both direct and BD-RIS-aided links, whereas the remaining UAVs are served through direct links only. We formulate a network sum-rate maximization problem that jointly optimizes UAV trajectories, RSMA common-rate allocation, beamforming, BD-RIS scattering matrices, and UAV-BD-RIS association. The resulting mixed-integer non-convex problem is solved by using deep deterministic policy gradient (DDPG) algorithm, equipped with a constraint-preserving action layer that generates reciprocal and lossless BD-RIS responses in closed form and handles the binary association inside the environment transition. Simulation results, averaged over three training seeds under a physically consistent channel model with blocked direct links near the BD-RISs, show that the proposed design attains a sum rate comparable to SAC and higher than twin delayed deep deterministic policy gradient ( TD3), proximal policy optimization (PPO), and deep Q-network (DQN), exceeds per-slot alternating optimization with fixed trajectories at about 1/65 of its per-slot runtime, and benefits mainly from RSMA, while the additional BD-RIS gain is modest.