Demand-Adaptive Virtual Base Station with UAV Swarm Distributed MIMO

Abstract

High-frequency wireless communication requires flexible coverage to cope with blockage and spatially varying traffic demand. This paper proposes a demand-adaptive virtual base station architecture consisting of a ground base station and UAV swarms forming distributed MIMO relays. The architecture includes two mechanisms: (i) virtual service-area partitioning based on spatial demand and (ii) adaptive UAV swarm formation that constructs distributed MIMO arrays for demand hotspots. For validation, we employ densitybased clustering and multi-agent reinforcement learning for UAV deployment. Performance is evaluated using equivalent virtual MIMO capacity under ray-traced propagation. Results show that the proposed method achieves near-optimal capacity compared to a genetic algorithm baseline while maintaining scalability and adaptability. The framework can be extended to time-varying traffic via sequential adaptation.