An improved K-means and DPC-empowered clustering approach for efficient routing in the FANET

With the rapid expansion of flying ad-hoc networks (FANETs), efficient routing algorithms have become crucial to ensure reliable and optimal data communication among Unmanned Aerial Vehicles (UAVs). In response to this demand, this paper introduced a novel approach called IKAD-FANET, a clustering-based routing algorithm for FANETs. The proposed algorithm combined the strengths of K-Means clustering and density peak clustering (DPC) to enhance routing performance. Specifically, K-Means was used to create initial clusters and select preliminary cluster heads based on spatial proximity, while DPC further refined cluster head selection by considering node density, ensuring stable and well-distributed clusters. The IKAD-FANET algorithm comprised two key stages. In the first stage, an enhanced K-Means algorithm was used to perform clustering, establishing cluster heads responsible for coordinating and managing UAVs within their clusters. This clustering process improved overall network organization and facilitated efficient data routing. In the second stage, the DPC algorithm calculated the density of nodes around each cluster head. This density information enabled a more intelligent and optimized assignment of nodes to their neighboring clusters. As a result, nodes were effectively distributed across the network, improving communication efficiency and resource utilization. Extensive simulations were conducted to evaluate the performance of IKAD-FANET, comparing its results to three existing methods: Q-FANET, CACONET, and GWOCNET. The simulation results demonstrated that the IKAD-FANET approach provided significant advantages over the other three methods in all evaluated criteria, outperforming them in terms of network lifetime, routing overhead, end-to-end delay, energy consumption, computational overhead, packet delivery ratio (PDR), cluster head lifetime, number of clusters, and clustering efficiency.