Vehicular sensing networks

Abstract

The increasing integration of sensors and wireless communication devices into highly mobile platforms such as automobiles makes vehicular sensing networks one of the most promising platforms for many applications. The performance of applications such as traffic reporting, environmental monitoring and distributed surveillance can be improved by using the new techniques developed in vehicular sensing networks. Several distinct features such as highly mobile and predictable movement patterns make vehicular sensing networks different from general computer networks. Therefore, unique solutions for vehicular sensing networks are necessary. In this thesis, we show our efforts on three aspects of vehicular sensing networks: efficiency, security and privacy. The efficiency issue is most critical when there is no central infrastructure or when vehicle-to-infrastructure (V2I) communication bandwidth is a precious resource. Keep uploading every piece of sensor data to a remote server is obviously inefficient. Local data aggregation is required to reduce the communication cost and improve the efficiency. However, how and when to perform the aggregation is not trivial. In this work, a GeoCache concept and Boomerang anchoring protocol are proposed to address this issue. Our work in security is focused on building secret keys for both vehicle-to-vehicle (V2V) and the V2I communication modes. Many applications require secure V2V and V2I communications and two sets of secret keys need to be created independently. Based on the special characteristics existing in vehicular sensing networks, we develop two key agreement algorithms to achieve the target. In many of the applications developed within vehicular sensing networks, GPS data has to be submitted to the central server continuously, which results in serious privacy violation. We have three contributions under the privacy domain. First, we develop a privacy preserving algorithm which protects user privacy without filtering too many location traces. Second, we study the possible privacy leakage due to the detailed location information available. At last, we extend the privacy system design under the assumption that there is no location proxy server and the information available is limited.