An Intelligent Quality of Service Architecture for Information-Centric Vehicular Networking

Abstract

Information-Centric Vehicular Networking (ICVN) is a paradigm shift in vehicular communication which implements self-content management to eliminates issues related to the current host-based IP network. Due to the exponential growth in demands of multimedia services, current vehicular networks face several challenges to support mobility with optimal Quality of Service (QoS). To achieve the high QoS performance, an intelligent vehicular network service with dynamic control mechanisms is necessary. This motivates the development of dynamic content management that implements an intelligence architecture. The proposed intelligent architecture comprised of two primary stages: classifications and discovery. In the first stage, a classifier system categorizes the user’s content request and the second stage presents an adaptive forwarding path discovery towards the nearest content provider. Here, a Rule based Evolutionary Systems (RES) agent performs exploration and exploitation to discover every possible forwarding path in the heterogeneous network. The agent then performs the discovery action that is guided by the variance introduced in the Reinforcement Learning (RL) policy. The simulation results confirms the suitability and scalability of the proposed architecture, particularly in reducing data packet delivery time, increasing data transfer rate, improve interest success rate, and lower the network traffic by 70%, 28%, 24% and 65% respectively.