Dr. Antonios Kontsos, Drexel University
This event is part of the Decision Sciences Seminar Series series.
Location:
Gerri C. LeBow Hall408
3220 Market Street
Philadelphia, PA 19104
Registration Option:
In the near future, humans, machines and the environment will function much closer to one another. The transformative step towards such “convergence” is the understanding and harnessing of complexity in the architecture, operation and performance of systems, as well as systems of systems across time and length scales. In this context, this talk focuses on current efforts of my research group in the direction of establishing data-driven intelligent systems. This term is inclusive with an emphasis on “systems”, as it is the most generic way to describe a myriad of assemblies, both physical and virtual, that are complex in nature and are tasked to perform under complex conditions. To form such synergies, sensing, data analysis and decision making processes are presented to demonstrate how models and algorithms could be created to achieve the goal of designing such data-driven intelligent systems. To describe the proposed research strategy and its applications, the case of cyberphysical systems is selected. First, the role of sensing is described as the necessary step to monitor the behavior of a complex system and collect data that could then be used for further understanding. Examples shown include materials and structures found in the built environment with emphasis on aerospace and civil applications. Then, the use of machine learning and data post-processing methods is explained as platforms to understand two elements: the current state of the system and when changes to it occur. Finally, the way such data could be used in mathematical or computational models both physics-based and phenomenological to predict the future state of such systems is described.