Visible to the public CRII: CPS: Towards Reliable Cyber-Physical Systems using Unreliable Human SensorsConflict Detection Enabled

Project Details
Lead PI:Dong Wang
Performance Period:05/01/16 - 04/30/19
Institution(s):University of Notre Dam
Sponsor(s):National Science Foundation
Award Number:1566465
446 Reads. Placed 450 out of 803 NSF CPS Projects based on total reads on all related artifacts.
Abstract: A growing number of Cyber-Physical Systems (CPS) domains, such as environment, transportation, energy, and disaster response, involve humans in non-trivial ways. Humans act as sensors in these scenarios when they contribute data (either directly or via sensors they own) that a CPS application can use. Using humans as sensors (commonly known as social sensing or crowdsensing) is an emerging paradigm, which provides unprecedented opportunities to sense the physical world in an inexpensive, versatile and scalable manner. However, these benefits are based on the assumption that the human-sensed data are reliable, but this is not always the case. In order for social sensing to become a viable component in CPS feedback loops, there is a critical need to understand the correctness of collected observations from unreliable individuals. This challenge is referred to as reliable social sensing. The objective of this project is to develop a new Reliable Social Sensing Model (RSSM) and system prototype, which enables correct reconstruction of states of physical environment from unreliable human sensors. This project leverages and innovates techniques in estimation theory and CPS to fill a critical gap in the rigorous analysis of human-sensed information, thereby providing a reliable social sensing component to build robust CPS with humans-in-the-loop. This project contains three key components. First, a RSSM will be developed to formally reason about the correctness of collective human observations and accurately assess the quality of analysis results. Second, a new reliable social sensing system prototype will be built to integrate the proposed RSSM with the state-of-the-art data processing techniques to handle different types of human sensed data. Third, by evaluating the proposed model and system through a real world social sensing application, the project will effectively validate the correctness of the RSSM and provide new insights into modeling humans as sensors for future research. The success of the project and follow-up work inspired by it could lead to a paradigm shift in CPS with human-in-the-loop by explicitly incorporating rigorous accuracy assessment into the development of new theories, systems and applications that rely on the collective observations from massive human sensors. The proposal is timely due to the increasing interests in social networks, big data, and human-in-the-loop systems, as well as the proliferation of computing artifacts that interact with or monitor the physical world. This research project will also contribute to the curriculum of CPS and Social Sensing courses, and will engage undergraduate and graduate students in STEM disciplines and from underrepresented groups.