Enhanced K-means-type Clustering Algorithm with Seeding Constraints for the VANET

Abstract

This paper considers a cluster-based relay vehicle selection scheme that adjusts the K-means-type algorithm to find real center mean vehicles under a restricted seeding range. With the help of the K-means-based VANET seeding principle, a local minimum cluster center is proposed, together with a rigorous proof by means of Lagrange multipliers. Unlike other existing works, relay vehicles in this correspondence are categorized into centroids and connecting vehicles to minimize the number of relay vehicles used and maximize the broadcasting power efficiency. To evaluate the system performance, we use different metrics to accommodate the realistic V2V scenario with NLOS signal dissemination, and empirical results show that both the algorithms themself and the system performance by implementing proposed algorithms are in the advantageous position over exsting appoaches.