Mobile Crowd-sourced Data Fusion and Urban Traffic Estimation

Abstract

Urban traffic estimation is one of the critical tasks for intelligent transportation systems (ITS). To estimate traffic condition, accurately and timely traffic data must be sensed frequently at every location around the city utilizing multimedia data fusion and analytics. This paper proposes a novel approach to urban traffic data collection and analysis leveraging crowd-sourced data from drivers and mobile users. Concretely, we have proposed solutions for mobile crowd-sourced data fusion to which just the right traffic data is collected automatically by GPS modules equipped in mobile devices. In addition, mechanisms for data validation and analytics for traffic estimation have been devised. Consequently, a mobile application is developed and provided to public users so that they can conveniently collect and share traffic data to the system. Besides, users can access traffic information and ITS services such as routing recommendation freely. The proposed system has been deployed for a real-world application in Ho Chi Minh City (HCMC), the largest city in Vietnam. Experimental results from real-field data confirm the feasibility, effectiveness and efficiency of the proposed approaches.