Developing Advanced Privacy Protection Mechanisms for Connected Automotive User Experiences

Abstract

The transportation industry is experiencing an unprecedented revolution. This revolution is being led by the rapid development of connected and automated vehicle (CAV) technologies together with cloud-based mobility services featured with huge amount of data being generated, collected,and utilized. This big data trend provides not only business opportunities but also challenges. One of the challenges is data privacy which is inherently unavoidable due to the information sharing nature of such mobility services and the advancement in data analytics. In this thesis, privacy issues and corresponding countermeasure that related to connected vehicle landscape are comprehensively studied. First of all, an overview of the landscape of emerging mobility services is provided and several typical connected vehicle services are introduced. Then we analyze and characterize data that can be collected and shared in these services and point out potential privacy risks. In order to protect user privacy while ensuring service functionality, we develop novel privacy protection mechanisms for connected automotive user experiences. Specifically, we consider the whole life cycle of data collection and sharing. To support privacy preserving data collection, we design fine-grained and privacy-aware data uploading policies that ensure the balance between enforcing privacy requirements and keeping data utility, and implement a prototype that collects data from vehicle, smartphone, and smartwatch securely. To support privacy preserving data sharing, we demonstrate two kinds of risks, additional individual information inference and user de-anonymization, during data sharing through concrete attack designs. We also propose corresponding countermeasures to defend against such attacks and minimize user privacy risks. The feasibility of such attacks and our defense strategies are evaluated with real world vehicular data.