UAV-assisted Semantic Communication with Hybrid Action Reinforcement Learning

• DOI: 10.48550/arxiv.2309.16713
• Publication Date: 2023-01-01
• Authors: Peiyuan Si; Jun Zhao; Kwok‐Yan Lam; Qing Yang
• Keywords: Benchmark (surveying)

In this paper, we aim to explore the use of uplink semantic communications with assistance UAV in order improve data collection effiicency for metaverse users remote areas. To reduce time while balancing trade-off between reconstruction quality and computational energy cost, propose a hybrid action reinforcement learning (RL) framework make decisions on model scale, channel allocation, transmission power, trajectory. The variables are classified into discrete type continuous type, which optimized by two different RL agents generate combined action. Simulation results indicate that proposed can effectively efficiency under parameter settings outperforms benchmark scenarios.