Comparative Analysis of Clustering Techniques for Deployment of Roadside Units

Today with the ever-growing demand of the internet and every second the transition to new technology, in-vehicle system also requires up-gradation. In this study, finding optimal positioning of roadside in vehicular Ad hoc Network (VANET) has been explored using Artificial Intelligence, as it is transforming every domain to a new level. Machine Learning can help us in predicting the optimal position of Roadside unit using the volume of vehicles and via verifying the longitude and latitude of the traffic vehicle. Various clustering techniques K-Means, Mean_Shift, Density-Based Spatial clustering of Application with Noise, Expectation_Maximization clustering (GMM) and Agglomerative_Hierarchical clustering has been applied on vehicle data consisting of longitude, latitude and volume of the taxi. Data was collected from NYC taxi (New York) from January 2016 to June 2016. Our results shows that machine learning provide excellent results in terms of position predictions.