|
This project investigates how privacy can be used to inform the design and management of future data sensing systems. Networked systems that collect data about individuals will play an increasingly important role in our lives, with applications including industrial monitoring and control, "smart" homes/cities, and personalized health care. These systems will gather private information about individuals, which creates many coupled engineering challenges. This work seeks to understand the interplay between two of these: managing the massive amounts of data that must be collected and protecting the privacy of individuals in the system. For applications or services that rely on populations of individuals, reducing the amount of information transmitted can save bandwith while enhancing privacy. The objectives of this work are to use ideas from data privacy technologies and wireless resource management techniques to jointly manage privacy and bandwidth in wireless sensing systems.
The technical objectives of this project involve reformulating distributed data collection and estimation problems under privacy constraints in wireless network settings. Differential privacy gives a framework for quantifying the privacy risk of different strategies for compression and data reduction, leading to a privacy-bandwidth-quality tradeoff. At the system level, total privacy risk is a resource that can be managed; this project uses that insight to design novel privacy-allocation schemes. Ultimately, this project seeks to formulate new tradeoffs between privacy and quality-of-service that can generalize to other networking problems. To see how these approaches work in practice, this work involves prototyping and testing the methods in a wireless networking testbed.
|
TWC: Small: PERMIT: Privacy-Enabled Resource Management for IoT Networks