Managing the Movement of Virtual Vehicles in Cloud Environment Using Reinforcement Learning Algorithm and Fuzzy Logic

Abstract

Traffic monitoring methods have remarkably high costs. Hence, using intelligent vehicles can reduce the costs related to infrastructures such as equipment application costs, system operation and its maintenance. Inter-vehicular communication provides the opportunity in the cloud environment for enhancing communication area. In this paper, distribution of a large volume of information by roadside units to the vehicular network is inspected based on effective vehicle-to-vehicle communication in cloud environment. Given the nature of vehicular networks, the topology of these networks changes rapidly. Therefore, a distributed pattern with simple protocols is taken into consideration, rather than focusing on centralized structures. By applying no-rate coding in roadside units and the utilization of vehicles as information carriers, an effective method is presented for distributing information to all nodes (even to nodes belonging to separate clusters) within the network. Furthermore, fuzzy logic, as a precise system for imprecise and approximate conditions in vehicular networks, and reinforcement learning, as an appropriate algorithm for dynamic environments, for managing the movement of vehicles are used. As a result, network efficiency is enhanced and data transmission delay in the network is reduced.