Georgia Tech Professor Yan Wang's NSF-funded (May 2017) research centers on probabilistic design of systems of Cyber-Physical systems (CPS), which are mechanisms and components with highly integrated functions of sensing, computing, communication, and control. These systems are interconnected and form the Internet of Things (IoT), with advanced capabilities to sense the environment, exchange information, and interact with humans. Such systems are becoming increasingly important in modern society. Future home and office appliances, manufacturing equipment, robots, medical implants and devices, and automobile are examples of CPS, which help establish the infrastructure of “Smart X” (smart home, smart office, smart medicine, smart city, etc.). Because of their intertwined connections with human beings and human society, they are also referred to as cyber-physical-social systems. Given the rising importance of CPS, improvements in their design can have major impacts on society. One challenge in designing such systems is to ensure the system is resilient, meaning that it can recover from major disruptions when a portion of the system fails to function or communicate. Another challenge is how to take social behaviors into consideration in their design so that people will feel comfortable to use such systems because they constantly collect and share our information. The goal of this research is to understand how to systematically design networked systems of cyber-physical systems such that they are resilient and trustable with reduced risks. This will include new techniques for measuring key properties of complex systems as well as techniques to design them with better performance.
For years, production managers at Mueller Inc. would stay late in the day or arrive in the wee hours of the morning to sketch out plans on paper for cutting parts out of giant rolls of steel. Since everything was done by hand — from making pattern plans to inputting that information into the cutting machines — getting a head start was essential.
“Few pieces of their equipment communicated with their software,” said Andrew Dugenske, principal engineer and director of the Factory Information Systems Center at Georgia Tech. “People had to enter data by hand, which was very labor intensive and error prone.
” Managers at the Texas-based manufacturer of metal buildings knew there must be a better way and enlisted factory technology researchers at Georgia Tech to help. The team built a system to link those cutting machines to computers across a network in the factory so production managers could automate the process and monitor the system with more confidence.
The project reduced Mueller’s scrap metal, increased production, and saved significant labor costs. “This is a great example of how the Internet of Things for manufacturing can provide significant value to manufacturers,” Dugenske said.
Scientists across the Georgia Tech campus are working on every aspect of a looming new reality: autonomous vehicles sharing the road with human drivers and revolutionizing how we travel by car. Those researchers include the School of Civil and Environmental Engineering’s Michael Hunter, director of the federally funded National Center for Transportation Systems Productivity and Management and the state-funded Georgia Transportation Institute. He uses computer models to study the management and operation of our future roadways, and he has identified a few of the issues policymakers and drivers will confront when self-driving cars travel our highways and bi-ways. Among them: how autonomous car could actually disrupt traffic flow and lead to surface-street bottlenecks.
Story can be found here: http://ce.gatech.edu/taxonomy/term/15
Dr. Jon G. Duke, M.D., joins the Georgia Tech Research Institute (GTRI) as Georgia Tech’s director of Health Data Analytics. Duke previously served as director of Health Analytics and Advanced Text Mining at the Regenstrief Institute at Indiana University. While at Regenstrief, he also lead the Drug Safety Informatics Lab as well as a 5-year partnership with Merck & Co, which conducted more than 45 projects involving at least 70 faculty and staff.
Duke will lead GTRI’s initiative to improve human health through better capture, interpretation, and applications of data. This effort will incorporate a spectrum of expertise including machine learning, natural language processing, high-performance computing, sensors, cybersecurity and health data interoperability.
While applying advanced technology, these efforts will manifest through real-world projects supporting not only research environments but health care systems, government and industry partners, and community collaborators.
Duke’s previous work focused on advancing techniques for conducting research through structured, unstructured and patient-generated health care data, with applications spanning research, quality and clinical domains.
His areas of expertise include the following: Big data analytics and natural language processing in health care Structured and unstructured clinical phenotypes Drug safety and clinical decision support UI/UX (user-interface/user experience) design in health IT applications Academic-industry collaborations
Over the last several years, Duke has directed more than $21 million in data research for industry and government sponsors. He has worked to expand on strategies for capturing better health care data, streamlining insights for stakeholders and delivering effective data-based interventions.
In 2014, Duke helped found the Observational Health Data Sciences and Informatics (OHDSI, pronounced “Odyssey”) program, which aims to develop open-source solutions to deliver value in health data through large-scale analytics.
Duke received his bachelor’s in 1994 from Emory University, and his M.D. from Harvard Medical School in 2000. He completed his internal medicine residency with Brigham and Women’s Hospital in Boston in 2003. In 2010, he earned a master degree in human-computer interaction from Indiana University. During this same time (2008 to 2010) he had a Fellowship in Medical Informatics with the Regenstrief Institute.
Board certified in internal medicine since 2003, Duke served as an adjunct professor of medicine, an adjunct professor of informatics and an adjunct professor of knowledge informatics and translation at the Indiana School of Medicine from 2010 to 2014. He was a resident clinical instructor at Harvard Medical School from 2000 to 2003.
In addition to co-founding the OHDSI Collaborative, Duke is a member of the Health Information and Management Systems Society, the American Medical Informatics Association and the American College of Physicians.