Intelligent Reward-Based Data Offloading in Next-Generation Vehicular Networks

Abstract

A massive increase in the number of mobile devices and data-hungry vehicular network applications creates a great challenge for mobile network operators (MNOs) to handle huge data in cellular infrastructure. However, due to fluctuating wireless channels and high mobility of vehicular users, it is even more challenging for MNOs to deal with vehicular users within a licensed cellular spectrum. Data offloading in the vehicular environment plays a significant role in offloading the vehicle’s data traffic from congested cellular network’s licensed spectrum to the free unlicensed WiFi spectrum with the help of roadside units (RSUs). In this article, an intelligent reward-based data offloading in the next generation vehicular networks (IR-DON) architecture is proposed for dynamic optimization of data traffic and selection of intelligent RSU. Within the IR-DON architecture, an intelligent access network discovery and selection function (I-ANDSF) module with $Q$ -learning, a reinforcement learning algorithm is designed. The I-ANDSF is modeled under a software-defined network (SDN) controller to solve the dynamic optimization problem by performing an efficient offloading. This increases the overall system throughput by choosing an optimal and intelligent RSU in the network selection process. The simulation results have shown the accurate network traffic classification, optimal network selection, guaranteed quality of service, reduced delay, and higher throughput achieved by the I-ANDSF module.