Enhancing V2V Communication in Autonomous Vehicles with Novel Clustering and Fuzzy Logic Framework

Autonomous vehicles rely on a clustering-based packet forwarding framework to communicate with one another in infrastructure-less environments. This framework typically involves clustering the vehicles into groups using the K-Means algorithm, identifying each group’s cluster heads (or leaders), and achieving inter-cluster communication through cluster heads. Since K-Means cannot handle arbitrary shapes and dynamic traffic conditions, the current clustering-based packet forwarding framework suffers from poor packet delivery ratios and network congestion. To tackle these problems, this paper proposes a novel density-based clustering algorithm named moving vehicle clustering (MVC) to cluster the vehicles in a network and identify the intelligent cluster head (CH) in each cluster with a novel fuzzy logic solution. The proposed framework achieves stable communication and optimized packet transactions. The experiments conducted using the SUMO and NS-3 network simulator demonstrate that the proposed framework significantly improves packet delivery ratios (over 85%), reduces delay, lowers queue sizes, and increases throughput.