AI-Enabled IoT Framework for Smart Traffic Surveillance and Communication

The rising increase in urban population and growing demand by passengers for seamless mobility requires efficient operation of road transport applications. The current traffic surveillance systems in most Nigeria cities are inefficient, costly, and characterized by errors thereby requiring the deployment of smart solutions. This paper proposes an AI-driven internet of things (IoT) solution to the present traffic flow management crises. We provide preliminary workflow activities for achieving our objective by designing a data dictionary of input features/parameters that impact traffic flow management. Specifications of the data dictionary assisted the simulation and implementation of a congestion prediction system modeled after four road networks within the Uyo Metropolis of Akwa Ibom State, Nigeria. Results of a deep learning model evaluation prove the success of our implementation with satisfactory values of root mean squared error (0.1332), receiver operating curve-area under the curve (0.9293) and accuracy (0.8668). A full-scale study including the implementation of our proposed framework is expected in future research.