Iterative greedy user clustering algorithm for D2D–relay in vehicular communication systems

Abstract

Device-to-device (D2D) communications help in improving the performance of wireless services in cellular networks via enabling cooperation among mobile users. Further, compared to traditional pairwise cooperation, cluster-wise cooperation can achieve even higher spectral and power efficiency. D2D clustering certainly can be a great supporting technology for the future of intelligent transportation. In this study, the authors investigate the D2D–relay clustering network in the vehicular scenario and propose a new clustering scheme named iterative greedy user clustering (IGUC). By exploiting those vehicles with better channel state, system resources can be saved. Here, they mainly focus on the question which is by applying D2D–relay clustering among vehicles, how much bandwidth could be saved in the uplink. This problem is mathematically formulated into a non-linear binary programming problem with the goal being minimum spectrum usage. Due to the NP-hardness, the low-complexity suboptimal algorithm IGUC is introduced to tackle it. Simulation results show that compared to the non-cooperative uplink system, IGUC saves a considerable proportion of the spectrum. Furthermore, clusters in denser and further areas tend to produce a higher gain. Influence of the fast time-varying characteristic of vehicular wireless channels is also investigated, upon which an adaptive switch mechanism is proposed to further facilitate IGUC.