Traffic flow Prediction at Signalized Road Intersections: A case of Markov Chain and Artificial Neural Network Model *Isaac O. Olayode Mechanical and Industrial Engineering Technology University of Johannesburg South Africa olayode89@gmail.com Lagouge K. Tartibu Mechanical and Industrial Engineering Technology University of Johannesburg South Africa ltartibu@uj.ac.za Modestus O. Okwu Mechanical and Industrial Engineering Technology Universty of Johannesburg South Africa okwu.okechukwu@fupre.edu.ng Abstract—Traffic congestion is a prevailing problem globally, which threatens the wider community, especially in developing countries. The Markov chain model (MCM) is a widely acknowledged and applied method used in traffic modelling, planning, and road traffic control systems. Classical techniques like MCM have been used to reduce vehicular flow and traffic congestions. Nowadays, artificial intelligence techniques have been recognized for solving traffic congestions and multivariate problems. The application of ANN in traffic flow prediction performance yielded positive results. The present study dwells on a comparison between the Markov Chain Model and artificial neural network model for predicting traffic flow of vehicles at signalized road intersections. Analysis of dataset collected at Mikros traffic monitoring (MTM) firm, with vehicular speed and distance as input variables and time as output, gave a good performance with root mean square error (RMSE) of 0.0025 and coefficient of determination (R 2 ) of 0.96417. The ANN model was adjudged capable of modelling traffic flow at road intersections. Keywords—Artificial Neural Network (ANN), Traffic congestion, Markov chain model, Artificial Intelligence. I. INTRODUCTION In the past few years, developed and developing countries have experienced elevated infrastructural and road network development levels. This development is accompanied by an increase in demand for mobility in urban cities [1]-[5], resulting in perennial problems such as high traffic demand mirrored by severe traffic congestions in metropolises [6]-[8]. These traffic congestion problems have led to a lack of travel efficiency (longer travel times), an upsurge in fuel consumption [9], and air pollution [10] caused by carbon monoxide discharging from exhaust pipes of vehicles stuck in traffic. Transportation researchers have tried to come up with both conventional and artificial intelligence approaches to address traffic-related problems such as traffic jams [11], [12], urban planning [13], [14], traffic prediction [15]-[18], and urban road system design. There remained a gap to be filled in tackling traffic congestion especially, traffic-related problems in road intersections. The application of an artificial neural network model in traffic congestion is not novel. However, many compelling questions have been raised by transportation researchers and academicians on the difference between a heuristics Markov model's strength and performance and a metaheuristics Artificial neural network model. This research is focused on comprehensive and extensive analysis of the prediction performance of traffic congestion of non-autonomous vehicles by using the ANN model and a heuristic Markov model. A comparative analysis of the application of ANN and a heuristic Markov model was done to predict traffic congestion of vehicles considering South Africa Roads. There are many ways to define “Transportation.” It is actually from two (2) Latin words, namely “trans” (which means going across) and “portare” (which means carry over). Transportation has to do with moving individuals from one place to another, including transporting and distributing goods and services from one place to another. Transportation can be an act of an individual or a systematic process, or a means of transporting or being transported by people/goods/services. It is a means of conveying goods and services through air, road, sea, or rail. The transportation can either be a private or public transportation system. Nowadays, transportation is very significant to the day-to-day activities of humans. Various modes of transportation are applied worldwide to transport individuals and freight from one location to another. Each type of transportation has its specific features of initialization and mode of operation. Having a perfect understanding of transportation subsystems is imperative to have a concise knowledge of transportation functions and the effects of the various stages of implementation and mode of transportation systems, dependent on the physical, biotic, and anthropic environments. Industrial revolution can be defined as the process of elevated change in a country's economy, primarily due to the inception of power-driven machinery. Another alternative definition is that the industrial revolution is the instantaneous rudimentary change in an industrial organization, the displacement or repudiation of the first, second, and third industrial revolution, which have all been replaced by the fourth industrial revolution. Industrial organizations play a significant role in the economic development of any country. These organizations also play an essential part in the driving force of any country’s economic growth. Any organizations' economic sustenance, albeit a small scale organization or a multinational organization, is all dependent on innovative thinking resulting from information and communication technology (ICT). The 287 2021 IEEE 12th International Conference on Mechanical and Intelligent Manufacturing Technologies 978-1-6654-1453-1/21/$31.00 ©2021 IEEE 2021 IEEE 12th International Conference on Mechanical and Intelligent Manufacturing Technologies (ICMIMT) | 978-1-6654-1453-1/20/$31.00 ©2021 IEEE | DOI: 10.1109/ICMIMT52186.2021.9476173 Authorized licensed use limited to: University of Johannesburg. Downloaded on July 14,2021 at 16:09:57 UTC from IEEE Xplore. Restrictions apply.