Research & Projects

  • Internet-of-Things (IoT) is transforming our lives with augmented everyday objects that have various levels of sensing, computing, actuating and networking capabilities. Currently, many IoT devices do not interact or exchange data with each other, they only offer limited information and services to users through specific mobile apps. Integration of services is mostly achieved in an ad-hoc manner using vendor-specific cloud APIs. Hence, many IoT services rely heavily on the cloud, which creates dependency on network connectivity and potentially increases latency for event-driven adaptations. Lastly, the lack of control and clarity on system adaptations often frustrate users.

    Status: Active | Tags: Cloud, BigData
  • Location data is important private information that enables many geosocial applications. Users have minimal control of their location privacy level in many of these applications, since they either share everything or nothing, especially after data is released. This project explores cryptographic approaches that support privacy-preserving location data sharing and proximity computation without any trusted-party and allow users to control the level of detail they are willing to reveal. We develop algorithms that provide robust spatial cloaking and constructions of homomorphic operations. Our spatial cloaking technique can significantly reduce unnecessary masking noise, and it is applicable to other cloaking solutions. 

    Status: Active | Tags: Homomorphic Encryption, Security, Privacy, Crypto
  • Driven by the proliferation of mobile devices there has been an increase in the provisioning of location-aware services to users. Location data of users is collected to construct trajectories that reveal mobility patterns. These patterns are utilized by organizations to serve a variety of purposes, including infrastructure and resource planning, environment protection, study of infectious disease dynamics, speedy evacuations during crises, and many others. However, exposure of mobility patterns is harmful to user privacy. Thus, there exists a conflict between service providers who wish to utilize users' location data for value added services and end users who prefer to have control over the handling of their location data.

    Status: Past | Tags: Privacy, Modeling, Prediction