Real-Time Data-Driven Framework for Route Prediction Based on VANETs

Abstract

With the advancement in technology, the implementations of smart cities and Internet of Things (IoT), etc. have enabled the integration of various devices that can communicate and facilitate in timely decision making. Similarly, traffic congestion is one of the very serious problem daily commuters are facing. In developed countries like USA, Germany etc. various sensors are used to gather real-time traffic information to analyze the traffic status. The traffic condition is communicated to all the commuters through the internet. The problem become more severe for developing countries where physical infrastructure and internet connectivity is not available at most of the highways. In this thesis, we have proposed a VCloud framework that enables route prediction through real-time data received from neighboring vehicles in ad hoc fashion in the absence of internet connection. The proposed scheme is implemented on embedded devices and evaluated in terms of energy and memory consumption. On the contrary to simulation/emulation based existing work, the proposed framework implemented on smartphone and evaluated on real-data. The data of live urban traffic is collected from all possible routes between two populous metropolitan locations. The route prediction through VCloud is analyzed using collected dataset in the absence of any road side units and internet connectivity. Moreover, data collected is found consistent with minimum variance by applying data quality measurement techniques.