Keynote Lectures

Evolving Systems and Their Automotive Applications
Dimitar Filev, Research & Advanced Engineering, Ford Motor Company, United States, United States

On Some Open-ended Challenges in Model-based Fault Management for Aerospace Systems: A Look Backwards and Forwards
Ali Zolghadri, University of Bordeaux - CNRS, France, France

The Science of Autonomy: A Holistic View at the Intersection of Learning, Control and Physics
Evangelos Theodorou, Georgia Institute of Technology, United States, United States

Evolving Systems and Their Automotive Applications

Dimitar Filev
Research & Advanced Engineering, Ford Motor Company, United States

Brief Bio
Dr. Dimitar Filev is a Henry Ford Technical Fellow at the Ford Research & Innovation Center, Dearborn, Michigan. He is conducting research in computational intelligence, AI and control, and their applications to vehicle systems, autonomous driving, and automotive engineering. Dr. Filev has published 4 books, over 200 journal articles and conference papers, and holds over 100 US and foreign patents. He is the recipient of the 2008 Norbert Wiener Award of the IEEE SMC Society and the 2015 Pioneer’s Award of the IEEE CIS Society. He received his PhD. degree in Electrical Engineering from the Czech Technical University in Prague in 1979. Dr. Filev is a Fellow of IEEE and a member of the NAE. He is past president of the IEEE Systems, Man, & Cybernetics Society (2016-2017).

Abstract
The emerging trend of increasing flexibility, adaptation, and autonomy of control and information systems is the driving force behind the evolving systems paradigm. Evolving systems are characterized with flexible model structure that adjusts to changes which cannot be solely handled by parameter adaptation. Evolving systems develop their structure and knowledge representation through continuous learning from data and interaction with the environment. They exploit synergies between two powerful concepts – real time data granulation and machine learning - with no limitations on the types of the model structure that may include regression models, neural networks, fuzzy, and/or stochastic models. Practical applications encompass a wide range of systems with variable parameters and structure, and multiple operating modes. This presentation provides an overview of the multiple facets of evolving systems theory and describes some of their automotive applications to adaptive process control, automated calibration, anomaly detection, driver state estimation, and fuel economy optimization.

On Some Open-ended Challenges in Model-based Fault Management for Aerospace Systems: A Look Backwards and Forwards

Ali Zolghadri
University of Bordeaux - CNRS, France

Brief Bio
Ali Zolghadri is a professor in Control & System Engineering at the University of Bordeaux, France. His research deals with model-based fault diagnosis, estimation and observation issues and fault-tolerant control & guidance methods. Over the last fifteen years, his focus has been on aerospace and flight-critical systems. More recently, he is conducting cross-domain research on fault management in interconnected, hybrid and distributed engineered systems using symbolic and abstraction-based methods. Between 2001 and 2015, he had been head of ARIA research team, IMS-lab, CNRS-Bordeaux University. He has authored and co-authored over 75 papers in leading international journals, about 130 communications in international conferences, one Springer book and 12 book chapters. He is a co-holder of 14 patents (French and US) in the aerospace field. He has been coordinator of a number of collaborative French, European and international research projects and actions in control and aeronautics. He is member of TC “Aerospace” and “Safeprocess" of IFAC and Council of European Aerospace Societies. He has served as a program committee member and international advisory board member for various international conferences and has given many invited keynotes, plenary talks and seminaries during international events. He has been member of executive committee of France’s Aerospace Valley Cluster and received an award for excellence in 2010 from the French Aeronautics and Space Foundation. He is the recipient of the 2016 CNRS Innovation Medal for outstanding scientific research with innovative applications in the technological and societal fields.

Abstract
Aerospace has always been a powerful engine of innovation. When we look to the future, it is not obvious to predict where the things are going but there is no doubt that the vector is pointed toward more autonomy and intelligence and that aerospace systems are becoming more distributed and more connected. Nevertheless, the maturity of technologies remains overriding and new unconventional technologies are only adopted if there is a proven need that cannot adequately addressed through conventional employed techniques. Yet, regulatory standards evolve as the industry matures and thanks to new innovative and disruptive technologies and digital transformation, a greater period of innovation is being opened to shape the future of aerospace. The talk will start with the current situation and a look backwards: about a half-century after the early academic works in model-based / data-driven fault management, there exists today a widening gap between advanced academic methods and real-world aerospace applications. The talk will attempt to highlight some of the main reasons for this situation. Next, the talk will argue that for the foreseeable future and given the predicted demands on aviation and aerospace industry, new distributed/cooperative model-based fault management methodologies will be required to enable paradigm shifts in future flight operational issues management. Solutions will arise from cross-domain research at the interface between system & control theory and computer science. Finally, the talk will discuss a future paradigm shift in civil aviation operations toward more autonomy in the cockpit, and some related challenges and opportunities.

The Science of Autonomy: A Holistic View at the Intersection of Learning, Control and Physics

Evangelos Theodorou
Georgia Institute of Technology, United States
https://pwp.gatech.edu/acds/

Brief Bio
Evangelos A. Theodorou is an assistant professor with the Guggenheim School of aerospace engineering at Georgia Institute of Technology. He is also affiliated with the Institute of Robotics and Intelligent Machines. Evangelos Theodorou earned his Diploma in Electronic and Computer Engineering from the Technical University of Crete (TUC), Greece in 2001. He has also received a MSc in Production Engineering from TUC in 2003, a MSc in Computer Science and Engineering from University of Minnesota in spring of 2007 and a MSc in Electrical Engineering on dynamics and controls from the University of Southern California(USC) in Spring 2010. In May of 2011 he graduated with his PhD, in Computer Science at USC. After his PhD, he was a Postdoctoral Research Fellow with the department of computer science and engineering, University of Washington, Seattle. Evangelos Theodorou is the recipient of the King-Sun Fu best paper award of the IEEE Transactions on Robotics the year 2012 and recipient of the best paper award in cognitive robotics in International Conference of Robotics and Automation 2011. He was also the finalist the best paper award in International Conference of Humanoid Robotics 2010 and International Conference of Robotics and Automation 2017. His theoretical research spans the areas of stochastic optimal control theory, machine learning, information theory and statistical physics. Applications involve learning, planning and control in autonomous, robotics and aerospace systems.

Abstract
Despite the recent advancements in Machine Learning and AI, there is an ongoing debate and skepticism in academia as well as in industry with regards to the applicability of AI algorithms to safety critical systems. This skepticism arises from the fact that scientific communities have different standards as to when something works and how often. Another reason for this skepticism is epistemelogical and has to do with the way how communities evolve and pose new scientific questions.

Motivated by this debate, in this talk I will present advancements in the area of autonomy that bridge the gap between the two scientific communities of control and AI. The underlying mathematical principles are at the intersection of stochastic control, statistical physics, machine learning and parallel computing. Inspired from these principles, new algorithms arise for perceptual control that can incorporate physics-priors for both control and perception. In addition, these perceptual control algorithms can be equipped with uncertainty quantification mechanisms for anomaly detection in safety critical. I will conclude this talk with future directions in the areas of safe AI, machine learning and control with applications to autonomy.