On the Design of Optimal One-Shot Distributed Combining for Cooperative Multi-UAV Systems

Abstract

Unmanned aerial vehicle (UAV) networks will become one of the key components of the future mobile communication systems, especially suitable for on-demand coverage extension and prompt capacity enhancement. However, UAV networks, detached from the ground infrastructure, are usually constrained for traffic-intensive network-wide signal processing. This paper tackles this issue by considering the multi-user signal detection at the UAV swarm, where the uplink signals of multiple ground users are cooperatively recovered. In particular, to reduce the signaling exchange and the coordination latency, a two-stage distributed minimum mean squared error (MMSE) detection is proposed for the UAV swarm, which merges the local MMSE detections at each UAV with one-shot weighted combining at a central station. The combining weights only depend on the long-term statistical channel state informations (CSIs) and therefore, are suitable for distributed network with limited signaling exchange. Numerical results show that the proposed distributed MMSE detection achieves similar performance to the centralized MMSE detection.