"A Systems Perspective of Data-driven Smart Cities"
The rise of low-cost sensors and the ubiquitous Internet have led to the unprecedented deployments of the Internet of things (IoT) devices. This has led to the emergence of smart and connected technologies that are revolutionizing domains such as healthcare, energy, and transportation, transforming the way we interact with information and physical systems. In this talk, I will describe the systems challenges that need to be addressed to fully realize the vision of smart cities and argue that well-understood design principles from traditional distributed systems and networking can be applied to derive principled approaches for designing next-generation smart IoT systems for smart cities. I will illustrate such an approach by first drawing upon principles of network congestion control and network fairness to design smart renewables that can modulate their power output in a grid-friendly manner. Analogous to TCP congestion control mechanism, our solar rate control protocol enables solar-based renewables to make decentralized decisions rate control decisions while optimizing the goodput of the smart grid. Next, I will describe how principles of virtualization and software-defined infrastructure from computer systems can be applied to virtualize IoT-based energy systems to flexibly multiplex and share larger systems across a set of users. To this end, I will present how virtualizing energy systems enable users to independently control and use their share of energy resources while providing greater autonomy for optimization and control. I will conclude my talk by summarizing my ongoing efforts in data-driven machine learning approaches for smart cities and present my future research vision for smart cities that will require a synergistic mix of systems and applied data science research, creating significant opportunities for interdisciplinary work.
Stephen Lee is a final year Ph.D. student in the College of Information and Computer Sciences at the University of Massachusetts Amherst, advised by Prof. Prashant Shenoy. His research area span several areas of computer systems, including distributed systems and cyber-physical systems, with an emphasis on domains such as smart cities, smart buildings, and transportation. His research is based on using principled computing systems methods and applied machine learning to solve data-driven problems in cyber-physical systems. He is a recipient of the UMass CICS Portfolio Distinction Award and the Tata Consultancy Service scholarship award.
Wednesday, April 10 at 12:20pm to 1:10pm
McAdams Hall, 114
821 McMillan Rd., Clemson, SC 29634, USA