ISR Researchers Take Best Paper at World Forum on IoT
By Josh Quicksall
Researchers from the School of Computer Science’s Institute for Software Research (ISR) were recently awarded the Best Paper at the 2016 IEEE 3rd World Forum on Internet of Things (WF-IoT).
Co-authored by Ivan Ruchkin, Selva Samuel, Bradley Schmerl, Amanda Rico, and David Garlan, “Challenges in Physical Modeling for Adaptation of Cyber-Physical Systems”, addresses current gaps in the modeling of artificial intelligence software that is heavily reliant on physical components in the real-world, such as hardware, people, and physical processes.
“The proliferation of ubiquitous technologies, like the Internet of Things, is driving computer scientists to fit increasing amounts of physics and traditional engineering into their approaches” says Ivan Ruchkin (pictured above), PhD student in Software Engineering and one of the paper’s co-authors, “This situation creates tension in classic CS approaches to modeling -- especially when it comes to making the system intelligent.”
Using the example of an experimental robotics platform, TurtleBot, the team illustrated three key challenges to modeling for adaption: selecting model formalism, obtaining physical models, and the use of physical models in adaptation.
From these challenges, the team suggests a shift in the overall perspective with regards to modeling which emphasizes the use of formalisms based on expressiveness, analyses, and expertise; the need for more guidance to connect model-building and model-value; and the importance of treating physical models as first-class entities.
The award comes as an honor and, Ruchkin notes, a bit of a surprise. The team however, is hopeful that their ideas will spur further exploration in this direction. “These issues are applicable not just to a particular robot, but broadly in many domains where autonomy and artificial intelligence are applied.”