Differentially Private Distributed Sensing
The growth of the Internet of Things (IoT) creates the possibility of decentralized systems of sensing and actuation, potentially on a global scale. IoT devices connected to cloud networks can offer Sensing and Actuation as a Service (SAaaS) enabling networks of sensors to grow to a global scale. But extremely large sensor networks can violate privacy, especially in the case where IoT devices are mobile and connected directly to the behaviors of people. The thesis of this paper is that by adapting differential privacy (adding statistically appropriate noise to query results) to groups of geographically distributed sensors privacy could be maintained without ever sending all values up to a central curator and without compromising the overall accuracy of the data collected. This paper outlines such a scheme and performs an analysis of differential privacy techniques adapted to edge computing in a simulated sensor network where ground truth is known. The positive and negative outcomes of employing differential privacy in distributed networks of devices are discussed and a brief research agenda is presented.
- Research Organization:
- Pacific Northwest National Lab. (PNNL), Richland, WA (United States)
- Sponsoring Organization:
- USDOE
- DOE Contract Number:
- AC05-76RL01830
- OSTI ID:
- 1344663
- Report Number(s):
- PNNL-SA-121595
- Resource Relation:
- Conference: IEEE 3rd World Forum on Internet of Things (WF-IoT 2016): Workshop on Ubiquitous Sensing and Actuation (UbSA) via the Internet of Things, December 12-14, 2016, Reston, Virginia, 216-221
- Country of Publication:
- United States
- Language:
- English
Similar Records
Integrate Light-Weight Deep Learning Tools with Internet of Things
Analyzing Data Privacy for Edge Systems