Series of Dept Research Seminars - "Some Results on Networked and Degraded Control Systems" (Date: 27 February 2017)

Our speaker Dr. Huadong Mo would like to share with us on "Some Results on Networked and Degraded Control Systems":  

Control systems appear in almost every aspect of science and technology. My research question focuses on how to understand, predict, control, and ultimately survive real-world control systems in an ever-changing world facing the global challenges of inherent disturbance, e.g. degradation, natural and human induced disasters. I will present three recently works in the field of network and degraded control systems: analysis, modelling, and control. (I) Robustness, a system’s ability to adjust its activity to retain its basic functionality when transmission delay and packet dropout occur, is a defining property of many networked control systems. I will show an analytical tool with which to identify state parameters of MAC protocol based transmission delays, helping us derive an effective model that accurately predicts these delays. The analytical results unveil the network characteristics that can enhance or diminish resilience, offering ways to prevent the collapse of and guiding the design of a wide-area power system resilient to transmission delays. (II) Increasing evidence shows that real-world components have different aging processes, and the real goal shouldn't just understand individual component’s degradation, but deciphering the dynamical interactions in the performance loss of entire system. A few aging components can cause cascading failures and catastrophic collapse of the entire system. I will show the general framework for analyzing the operation cost of distributed generation systems. The results have been surprisingly rich, and they differ so greatly from those of single or several components that they present a new paradigm. (III) We will know that we can describe attacker’s behaviors when we reach the point that we understand the preference attack time. I will present a data-driven model for attacker’s behaviors and the corresponding damage cost model for minimizing the unsupplied demand of a cyber-physical system. After the three aspects of current research, I will briefly describe my research topics, including traditional reliability analysis, optimal operations of networked and degraded control systems, real-time management of distributed generation systems and resilience increase of cyber-physical systems.

All are welcome and no registration is in need.