Publication

Abstract : Nowadays the number of cars with empty seats travelling a long distance are increasing on the roads. By bringing people, who are travelling to the same destination, in a single car can decrease the number of vehicles on the road and thereby reduce the pollution to a large extent. Web and mobile applications for carpooling are very common now which provide basic features like sharing of journey, user rating etc. However, all these applications could not provide efficient location tracking. In this paper, we propose an efficient mobile application for carpooling with some unique features like user location tracking and traffic anomaly detection. The location tracking is included as a security feature which enables the user to share his or her current location with their near and dear ones. Anomaly detection feature can reduce the unnecessary wastage of time during the journey by analyzing the pre reported traffic anomalies like processions, accidents, road works etc. reported by others and redirecting the vehicle through another best route. But there is a chance for false anomaly reporting possible. In order to make this feature foolproof, we propose a truth estimation technique using recursive EM algorithm in this paper. Recursive EM algorithm is an efficient streaming algorithm generally implemented in social sensing application as it can solve the estimation problem on a real time basis. This android based application consist of a user friendly interface through which user can create and manage trips, track location, report an anomaly, etc. Location tracking is implemented using Google Map API by collecting the GPS data from each client and feeding it into the database in real time.