Title: Leveraging Digital Twins and Adaptive Multi-Agent Deep Reinforcement Learning for Smart Spectrum and Resource Allocation in Open-RAN UAV-Enabled 6G Networks

Abstract: As wireless networks transition toward 6G, the vision includes a highly intelligent, Open-RAN-integrated infrastructure where unmanned aerial vehicles (UAVs) are essential for boosting coverage, strengthening resilience, and maintaining reliable connectivity for ground-based users. Nevertheless, the effective management of spectrum and resources within these rapidly changing UAV-supported ecosystems presents a significant hurdle, driven by nonlinear system dynamics, topology shifts caused by mobility, and strict requirements for low latency and energy efficiency. To overcome these obstacles, this study introduces a digital twin (DT)-enhanced adaptive deep reinforcement learning (DRL) framework designed to facilitate intelligent spectrum sharing and resource distribution among dispersed ground users. The intricate optimization challenge is split into two parts: optimizing UAV trajectories through particle swarm optimization (PSO) and handling dynamic spectrum, power, and association management using multi-agent DRL (MADRL). This hybrid strategy, powered by digital twins, fosters intelligent, context-sensitive decision-making and adaptive coordination among UAVs. Comprehensive simulations reveal substantial improvements in spectral efficiency, data throughput, and energy consumption, highlighting a promising route toward self-adapting, autonomous connectivity between 6G UAVs and ground users.

Source: arXiv Generated at: 2026-06-02 00:00:00 UTC