Intelligent Symbiotic Relay Selection Technique for 5G Networks

Abstract

The growing demand for bandwidth and spectrum has inspired the ongoing efforts to establish the future 5G network supporting vertical sectors such as cyber-physical systems (CPS). Cooperative communication is one of the requisite techniques to improve coverage, network capacity and reduce power consumption in the network. In this paper, a symbiotic two-phase intelligent transmission is considered. The first phase occurs between the source and the candidate relays, and involves the selection of a set of “reliable relays”. The second phase occurs between the reliable relays and the destination, and involves the selection of the “best relay” for transmission. Dynamic relay selection using k-means clustering is used to detect the most significant correlation between all the channel state information (CSI) attributes in the system. The work identified the reliable relays while reducing the number of relay nodes for the second transmission phase. Contextual scenarios are created with typical network configuration using three geographical locations Coventry, Birmingham and London. An experimental validation is done with Omnet++ environment for the scenarios of three geographical locations. A natural grouping of mobile users is carried out leveraging the relay capabilities. The results are validated using support vector machine (SVM) classification algorithm. Considering urban environment deployment of relay nodes, metrics such as signal-to-noise-plus-interference ratio (SINR), attenuation, signal to noise ratio (SNR), link quality, k-means clustering, accuracy, and root mean square error (RMSE) are investigated for the Direct-2-Direct (D2D) capable relays. It was observed that the proposed technique both outperforms the other fixed-parameter relay selection techniques and improves with larger datasets unlike the other techniques.