Machine Learning Approach on Traffic Congestion Monitoring System in Internet of Vehicles
Shridevi Jeevan Kamble, Manjunath R Kounte
Abstract
Traffic congestion is a major issue in urban cities leading to aggregated traffic. With the advancement in intelligent internet of vehicles, new technologies and protocols have been developed to predict the traffic congestion and utilize this traffic-related knowledge for congestion prediction and identification. This paper highlight the ML approach to identify traffic congestion based on multiple parameters such as hard delay constraints, the speed available through GSP vehicle trajectory. Here, we have used the Gaussian process in ML for prediction of traffic speed which uses 3 datasets i.e. training set, prediction set, and road sector data frame. ML can provide live traffic prediction in real-time, future traffic prediction and short-term traffic prediction on recent observation and historical data. In this paper using the data set, we have identified three different time slots for vehicle traffic congestion monitoring and evaluated the average speed of vehicles on the road sector during respective time slots.