Toward Location based Vehicular Networks in Urban Environments

Vehicular network have attracted tremendous amount of interests recently from both the academia and industry. In this paper, we first analyze a large set of data consisting of location information of about 8900 privately owned vehicles over a 3 months period of time. Compared to widely studied data set on taxis or buses, we show that privately owned vehicles exhibit significantly different characteristics, including dominant parking time, highly repetitive moving pattern, and concentrated visited places in urban environments. Motivated thereby, we propose a different type of vehicular networks that center its operations on places. We focus on visited places of the vehicles and resulting opportunistic yet relative stable contacts within the places. In these opportunistic networks, data exchange only happens in each place, such as a parking lot or street sides. While vehicles move from one place to another, on vehicle data will be carried as well and hence form a network composed of places and linked by moving vehicles.