On Passive Privacy-Preserving Exposure Notification Using Hash Collisions
Phillip Smith,Shamik Sarkar,Neal Patwari,Sneha Kasera
DOI: https://doi.org/10.1109/jiot.2024.3353255
IF: 10.6
2024-01-01
IEEE Internet of Things Journal
Abstract:Even as the COVID-19 pandemic drove advances in contact tracing and exposure notification systems, user privacy challenges continue to plague otherwise promising approaches to contain contagions. We propose a novel, scalable approach to address privacy in contact tracing that improves utility. We apply passive WiFi scan data using two metrics suitable for estimating contact between users. We support this with real world experimental data captured across a range of environments relevant to contact tracing. To preserve privacy, we leverage properties of truncated cryptographic hashes in an adaptation unique to contact tracing. This hash collision filter allows users to share information about potential contacts with a central server without revealing sensitive information. Using an aggressive threat model, including adversarial users and a malicious server, we share how this technique can improve utility while still providing strong security protections compared to other approaches using, for example, only Bluetooth (BT) or global navigation satellite systems (GNSS). Finally, we discuss a capability of this approach that allows notification for asynchronous co-location from past contacts.
computer science, information systems,telecommunications,engineering, electrical & electronic