A2T-Boost: An Adaptive Cell Selection Approach for 5G/SDN-Based Vehicular Networks

Abstract

Heterogeneous ultra-dense networks (HUDNs) are one of the key enabling technologies for the fifth-generation (5G) networks. They aim to provide high capacity, low installation cost, and distributed traffic loads. The cell selection is a challenging issue in HUDNs, due to the different characteristics of base stations (BSs) and the existence of a large number of them. Thus, the traditional cell selection scheme is not applicable in such a network. In this paper, a novel adaptive cell selection strategy is proposed, called adaptive two-tier based on adaptive boosting (A2T-Boost). It can adapt to the various characteristics of base stations, as well as the different movement features of mobile stations such as vehicles and pedestrian. It is a software-defined networking (SDN)/machine learning (ML)-based scheme. A real-world case is considered in the downtown of Los Angeles city. Simulation results demonstrate that A2T-Boost achieves high prediction performance and it outperforms other related schemes in terms of average number of handovers (HOs) by up to 50%. Moreover, it enhances the average achievable downlink sum-rates and network energy efficiency achieved by vehicles by up to 33.76%. Furthermore, the average packet delay is decreased using the proposed scheme by up to 12.87%.

Keywords

• 5G
• small cells
• SDN
• Los~Angeles
• machine learning
• HUDNs