2018-2019 ECE Distinguished Lecture Series
Edward Coyle, Georgia Institute of Technology
"Theory, Design, and Implementation of Large-Scale Wireless Sensor Networks: Combining Research and Education via a VIP Team"
Edward J. Coyle is the John B. Peatman Distinguished Professor of Electrical and Computer Engineering at the Georgia Institute of Technology and a Georgia Research Alliance Eminent Scholar. He is the Founder and Director of the Vertically Integrated Projects (VIP) Program, which integrates research and education by embedding large-scale, long-term teams of undergraduates in the research efforts of faculty and their graduate students. He is also the Director of the VIP Consortium, a set of 28 universities that have VIP Programs and work together to improve, evaluate, and disseminate it. Dr. Coyle was a co-recipient of the U.S. National Academy of Engineering’s 2005 Bernard M. Gordon Prize for Innovation in Engineering and Technology Education. In 1998, Dr. Coyle was elected a Fellow of the Institute of Electrical and Electronics Engineers (IEEE) for his contributions to the theory of nonlinear signal processing and he is a Distinguished Engineering Alumnus of the University of Delaware. He has received a number of other awards, including the 1997 Chester F. Carlson Award from the American Society for Engineering Education and the 1986 Paper Award from IEEE Signal Processing Society. His current research interests include systemic reform of higher education, signal and information processing, and wireless and sensor networks.
Large-scale wireless sensor networks that consist of thousands of battery-powered nodes offer unique opportunities to monitor crowds and civil infrastructure. But their large scale also presents many design challenges when the goal is the collection of data in an energy efficient manner but with a deadline by which a reliable global decision/estimate must be produced. In this talk, we first show how hierarchical clustering under application-dependent aggregation can enable energy efficient collection of data across the network. We then consider the challenge of using such an architecture when there are time constraints on the collection of the data. Such constraints imply that there is not enough time to collect data from every node in the network, so standard random-access protocols will not suffice. We thus propose and analyze a combination of random-access protocols and population splitting algorithms to gather only the most reliable local decisions and estimates during the time available.
The theoretical results described above have alternately been inspired by and informed the design, deployment, and operation of three sensor networks in the football stadium at Georgia Tech. Football games are attended by 50,000+ fans and staffed by 1000s of media, medical, and team personnel. They thus represent a unique opportunity for sensing fan activity, monitoring the stadium, and understanding RF spectrum usage. We give an overview of three of these systems, which have all been created by VIP teams at Georgia Tech, and show how they enable new opportunities for the entertainment of fans, the education of students, and the verification of research results.
Wednesday, May 1, 2019 at 3:30pm
Mitchell Hall, Auditorium
Mitchell Hall, University of Delaware, Newark, DE 19716, USA