Design and Implementation of Intelligent Traffic-Management System for Smart Cities using Roaming Agent and Deep Neural Network (RAD2N)

Abstract

In metropolitan areas, the exponential growth in quantity of vehicles has instigated gridlock, pollution, and delays in the transportation of freight. IoT is the modern revolution which pushes the world towards intelligent management systems and automated procedures. This makes a significant contribution to automation and intelligent societies. Traffic regulation and effective congestion management assist conserve many priceless resources. In order to recognize, collect and send data, autonomous vehicles are furnished with IoT powered Intelligent Traffic Management System (ITMS) having a set of sensors. Moreover, machine learning (ML) algorithms can also be employed to enhance the transportation system. Traffic jams, delays, and a high death rate are the results of the problems that the current transport management systems face. In this paper, an active traffic control for VANET is proposed which merges Roaming Agents (RA) with deep neural networks (DNN). The effectiveness of the DNN with RA (RAD2N) routing method in VANETs is evaluated experimentally and compared with the traditional ML and other DL routing algorithms. Several traffic congestion indicators, including delay, packet delivery ratio (PDR) and throughput are used to validate RAD2N. The outcomes demonstrate that the proposed approach delivers lower latency and energy consumption.