Publication Type:

Journal Article


Volume 10, p.2101-2106 (2015)



The proposed work suggests a model to sense various traffic anomalies like accidents, celebrations, protests, disasters etc., and thereby avoiding the traffic congestion in those areas so that the travelers can save their time and money by the automatic route suggested by the application. The system also proposes a method to check the authenticity of those detected anomalies with the help of social media. The application works in a distributed environment where an android application acts as the client, J2EE application as server and Hadoop in the back-end. The client side application contains basic navigation features along with an interface to report various traffic anomalies by the users. The proposed work maintains a twitter account to tweet the exact location and incident details to post automatically by the system. The authenticity of those reported incidents/anomalies are verified by the Recursive Expectation Maximization Algorithm. After ensuring the authenticity of the reported anomaly, all the users in that particular route will get intimation in advance. The system also suggests a best alternate path to the destination. MapReduce framework is used to process bulk amount of GPS data received during travelling with the help of Hadoop based infrastructure which is deployed in the backend. The system has been tested successfully using android and GPS location spoofing application.

Cite this Research Publication

P. K. Binu, Jisha R. C., Sai, A., and Salim, S., “A Hadoop Based Architecture Using Recursive Expectation Maximization Algorithm for Effective and Foolproof Traffic Anomaly Detection and Reporting”, vol. 10, pp. 2101-2106, 2015.