Clustering-based Multicast Scheme for UAV Networks

TL;DR

This paper proposes a clustering-based multicast scheme for UAV networks acting as aerial small base stations, reducing delay and signaling overhead by enabling UAVs to share packets within clusters instead of relying on retransmissions from the base station.

Contribution

A novel multicast scheme tailored for UAV networks using clustering and CSMA/CA, with analytical modeling via stochastic geometry and extensive simulation validation.

Findings

The scheme reduces delay and signaling overhead in UAV multicast communications.

Analytical models accurately predict the performance of the multicast scheme.

Simulation results confirm the scheme's superiority over traditional methods.

Abstract

When an unmanned aerial vehicle (UAV) network is utilized as an aerial small base station (BS), like a relay deployed far away from macro BSs, existing multicast methods based on acknowledgement (ACK) feedback and retransmissions may encounter severe delay and signaling overhead due to hostile wireless environments caused by a long-distance propagation and numerous UAVs. In this paper, a novel multicast scheme is designed for UAV networks serving as an aerial small BS, where a UAV experiencing a packet loss will request the packet from other UAVs in the same cluster rather than relying on retransmissions of BSs. The technical details of the introduced multicast scheme are designed with the carrier sense multiple access with collision avoidance (CSMA/CA) protocol for practicability and without loss of generality. Then, the Poisson cluster process is employed to model UAV networks to capture their dynamic network topology, based on which distance distributions are derived using tools of stochastic geometry for analytical tractability. Additionally, critical performance indicators of the designed multicast scheme are analyzed. Through extensive simulation studies, the superiority of the designed multicast scheme is demonstrated and the system design insight related to the proper number of clusters is revealed.