Improved Affinity Propagation Clustering for D2D Communication in 5G

Abstract

Fifth-generation mobile network(5G) is the latest cellular technology after 4G networks. The Use of a 5G network enhances the data rate due to more available bandwidth and advanced technologies. Among several other technologies, Device-to-Device (D2D) communication is one of the advanced technologies used in 5G networks for efficient data transmission. D2D technology represents direct communication between two devices without the assistance of a base station. The clustering algorithm is one of the technologies used mostly in D2D communication to handle dynamic devices. The clustering technique helps to group users with similar interests and reconstruct the network to achieve better performance in terms of throughput, spectral efficiency, power, energy consumption, etc. The Affinity Propagation (AP) clustering algorithm has differentiated itself from the other clustering algorithms by dynamically preparing the clusters and cluster heads. However, the other clustering algorithms require the number of clusters and cluster head information beforehand. Hence, this work focuses on improving the AP clustering algorithm to prepare the clusters and cluster heads in a better way to enhance the efficiency of D2D communication.