DescriptionAd hoc wireless networks have emerged as a solution to providing ubiquitous, on-demand connectivity without the need for significant infrastructure deployment. In this thesis we address the privacy problems in two types of emerging wireless ad hoc networks, namely sensor and vehicular networks.
Although the content of sensor messages describing ``events of interest'' may be encrypted to provide confidentiality, the context surrounding these events may also be sensitive and therefore should be protected from eavesdroppers. The source-location privacy problem occurs in sensor networks when adversaries use RF localization
techniques to perform hop-by-hop traceback of messages to the source sensor's location. Our work provides a formal model for this problem
and examines the privacy characteristics of different sensor routing protocols. In order to provide efficient and private sensor communications, we devise new techniques to enhance source-location privacy that augment these routing protocols.
Similarly, an adversary armed with knowledge of the network deployment, routing algorithms, and the data sink location can infer the temporal patterns and track the spatio-temporal evolution of a sensed event, by monitoring the packet arrivals at the sink. We introduce the temporal privacy problem for delay-tolerant sensor
networks, provide an information theoretic formulation and propose adaptive buffering to obfuscate temporal information from the adversary.
Sensor networks are also characterized by distinctive traffic patterns, wherein traffic mostly exists when events of interests occur. Due to the direct correlation between the type of event and
size of data generated by it, an adversary observing a traffic burst can infer information about the type of event simply from the observed message size. We formulate this traffic privacy problem in terms of information entropy, present a quantifiable means to measure traffic privacy and propose solutions to enhance it.
Vehicular ad hoc networks represent a promising new communication paradigm that can facilitate many new forms of automotive applications. We present a robust and efficient security and privacy
framework, for such networks, that uses identity-based cryptography. We show that our framework provides authentication, confidentiality, non-repudiation and message-integrity. Further, it supports scalable, user-customizable privacy through short-lived,
unforgeable, pseudonyms.