Statistical Analysis of Optimization-Based Clustering Scheme for Multi-UAV Networks

Abstract

The advancement in the widespread application of multi-UAV (Unmanned Aerial Vehicle) systems introduces certain design challenges in terms of communication and network stability. In this paper, addressing these design aspects, a clustering scheme is presented. The paper aims to improve the lifetime of a Flying Ad-hoc Network (FANET) considering the node distances, residual energy, delay and coverage as the fundamental elements for the election of the cluster heads and simultaneous formation of clusters. The clustering scheme is carried out by four different well-known metaheuristic techniques, i.e., Water Cycle Algorithm (WCA), Crow Search Algorithm (CSA), Firefly Algorithm (FA) and the Cuckoo Search Algorithm (CuSA). In addition, the topology of the network is maintained with periodic updating of the positions of the UAV nodes. Furthermore, the performance of these techniques is evaluated and discussed in terms of the time required for the cluster formation, network energy consumption, alive node analysis and finally, the network lifetime. Rigorous simulations are then carried out for different network areas with varying node densities. The comparative analysis demonstrates a significantly improved performance of the Crow Search based scheme for clustering over the other implemented techniques.