One of the major areas of research in the Coordinated Science Laboratory is control. In fact CSL was originally named “Control Systems Laboratory” when it was founded in 1951. From then on, CSL has become one of the leading research centers in the field of control (Ranked 2nd worldwide). This great success is due to the CSL’s unique collaborative environment that promotes interdisciplinary research.
The Decision and Control session in CSL Student Conference provides an opportunity for researchers in this area to gather and discuss the latest advances in the field of control. The session includes a keynote lecture by a distinguished speaker followed by several student talks presenting their current research. We invite students across ALL majors to present their work in the areas related to (but not limited to) control theory, optimization, game theory, and robotics.
Optimal Mass Transport and the Robustness of Complex Networks
February 23, 9:00-9:50, CSL B02
Today’s technological world is increasingly dependent upon the reliability, robustness, quality of service and timeliness of networks including those of power distribution, financial, transportation, communication, biological, and social. For the time-critical functionality in transferring resources and information, a key requirement is the ability to adapt and reconfigure in response to structural and dynamic changes, while avoiding disruption of service and catastrophic failures. We will outline some of the major problems for the development of the necessary theory and tools that will permit the understanding of network dynamics in a multiscale manner.
Many interesting networks consist of a finite but very large number of nodes or agents that interact with each other. The main challenge when dealing with such networks is to understand and regulate the collective behaviour. Our goal is to develop mathematical models and optimization tools for treating the Big Data nature of large scale networks while providing the means to understand and regulate the collective behaviour and the dynamical interactions (short and long-range) across such networks.
The key mathematical technique will be based upon the use optimal mass transport theory and resulting notions of curvature applied to weighted graphs in order to characterize network robustness. Examples will be given from biology, finance, and transportation.
Allen Tannenbaum is an applied mathematician and presently Distinguished Professor of Computer Science and Applied Mathematics & Statistics at the State University of New York at Stony Brook. He is also Investigator of Medical Physics at Memorial Sloan Kettering Cancer Center in New York City.
Tannenbaum has done research in numerous areas including robust control, computer vision, and biomedical imaging, having more than 500 publications. He pioneered the field of robust control with the solution of the gain margin and phase margin problems using techniques from Nevanlinna–Pick interpolation theory, which was the first H-infinity type control problem solved. He was one of the first to introduce partial differential equations in computer vision and biomedical imaging co-inventing an affine-invariant heat equation for image enhancement. Tannenbaum and collaborators further formulated a new approach to optimal mass transport (Monge-Kantorovich) theory. In recent work, he has developed techniques using graph curvature ideas for analyzing the robustness of complex networks.
His work has won several awards including IEEE Fellow, O. Hugo Schuck Award of the American Automatic Control Council in 2007 (shared with S. Dambreville and Y. Rathi), and the George Taylor Award for Distinguished Research from the University of Minnesota in 1997. He has given numerous plenary talks at major conferences including the IEEE Conference on Decision and Control of the IEEE Control Systems Society in 2000, the International Symposium on the Mathematical Theory of Networks and Systems (MTNS) in 2012, and the SIAM Conference in 2017.
Karthik Gopalakrishnan, MIT
Stability and control of switching network models: An air traffic delay example
February 23, 10:00-10:30, CSL B02
Network models have been proposed for a variety of systems ranging from social networks, electricity grids, financial institutions and so on. Most models presented in literature assume that the network topology, i.e. the nature of interaction between the nodal elements do not vary with time. In this talk, we will present an approach to model such switching network behaviour as a Markov Jump Linear System (MJLS) and analyse it. The talk is divided into three major parts. In the first part of this work, we formulate a MJLS model for the network dynamics. We illustrate the modelling approach with an air traffic delay network example, which will be used throughout the talk. In the second part, we discuss the notion of stability for a networked system described as a MJLS. Two notions of asymptotic stability are considered: Mean Stability and Almost Sure Stability, and necessary and sufficient conditions are derived. In the final part, we will motivate the new class of controllers that can be developed for switched networks. This lays the foundation for analysing switching-network dynamics and design optimal control and recovery strategies.
Karthik Gopalakrishnan is a PhD candidate in the department of Aeronautics and Astronautics at the Massachusetts Institute of Technology. His research focuses on the application of optimization methods and control theory to analyse networked systems, with an emphasis on air transportation. He is a recipient of the best paper award at the International Conference on Research in Air Transportation (2016) and a silver medal for his undergraduate academic performance from IIT Madras (2014).
Analysis, Identification, and Validation of Discrete-Time Epidemic Processes
February 23, 10:30-10:50, CSL B02
Models of spread processes over non-trivial networks are commonly motivated by modelling and analysis of biological networks, computer networks, and human contact networks. However, identification of such models has not yet been explored in detail, and the models have not been validated by real data. In this work, we present several different spread models from the literature and explore their relationships to each other; for one of these processes, we present a sufficient condition for asymptotic stability of the healthy equilibrium, show that the condition is necessary and sufficient for uniqueness of the healthy equilibrium, and present necessary and sufficient conditions for learning the spread parameters. Finally, we employ John Snow’s seminal work on cholera epidemics in London in the 1850’s to validate an approximation of a well-studied network-dependent susceptible-infected-susceptible (SIS) model. The validation results are surprisingly good, capturing the behaviour of the cholera epidemic from John Snow’s 1854 dataset quite well.
Philip E. Paré received his B.S. in mathematics with University Honours and his M.S. in Computer Science from Brigham Young University, Provo, UT, in 2012 and 2014, respectively. He is currently an ECE Ph.D. candidate at the University of Illinois at Urbana-Champaign, Urbana, IL and is a 2017-2018 College of Engineering Mavis Future Faculty Fellow. Philip grew up in Cambridge, MA. His research interests include the modelling and control of dynamic networked systems, model reduction techniques, and time-varying systems.
Muhammed O. Sayin
Secure Sensor Design Against Undetected Infiltration: Minimum Impact-Minimum Damage
February 23, 10:50-11:10, CSL B02
We propose a new defense mechanism against undetected infiltration into controllers in cyber-physical systems. To this end, we cautiously design the outputs of the sensors that monitor the state of the system. Different from the defense mechanisms that seek to detect infiltration, the proposed approach seeks to minimize the damage of possible attacks before they have been detected. In particular, controller of a cyber-physical system could have been infiltrated into by an undetected attacker at any time of the operation. Disregarding such a possibility and disclosing system’s state without caution benefits the attacker in his/her malicious objective and correspondingly can lead to severe consequences. Therefore, secure sensor design can improve security of cyber-physical systems further when incorporated along with other defense, e.g., infiltration detection, mechanisms. Furthermore, through secure sensor design, we also seek to impact the ordinary operations of the system at minimum while minimizing the damage due to inconspicuous, undetectable or difficult to detect, attacks. We, specifically, consider a controlled Gauss-Markov process, where the controller could have been infiltrated into at any time within the system’s operation. In the sense of game-theoretic hierarchical equilibrium, we provide a semi-definite programming based algorithm to compute the optimal linear secure sensor outputs that lead to both minimum impact and minimum damage on the ordinary operations of the systems. We also analyze the performance for various scenarios numerically and address the robustness of the algorithm against inaccurate perception of attack statistics.
Muhammed O. Sayin is currently pursuing the Ph.D. degree in Electrical and Computer Engineering from the University of Illinois at Urbana-Champaign (UIUC). He received the B.S. and M.S. degrees in Electrical and Electronics Engineering from Bilkent University, Ankara, Turkey, in 2013 and 2015, respectively. His current research interests include signaling games, dynamic games and decision theory, stochastic control, and cyber-physical systems.
Multi-Agent Learning for Coordinated Robotic Weed Killing
February 23, 11:10-11:30, CSL B02
This paper presents a learning approach for coordinated robotic weeding based on dynamic programming. The weeding problem is framed as a factored MDP, and a policy is learned asynchronously for several homogenous agents moving through a field environment modelled as a discrete grid representing one acre of crop land. This paper includes simulated experiments within a Weed World environment containing a dynamic weed-growth model. In these simulated experiments, the reactive policy algorithm was able to complete the weeding task, even in the presence of decreased observation radius and increased rate of weed-growth. This suggests the feasibility of this approach for the real-world weeding task and motivates further work.
Wyatt McAllister received his Bachelor’s Degree with Honours and Distinction in 2016 in the Department of Electrical and Computer Engineering at Illinois. He will complete his Master’s Degree by 2018 in Autonomy and Learning-Based Control and intends to pursue doctoral study. He is motived to create new technologies for applications in beneficial consumer infrastructure, and to research the electronics and the computing systems that comprise them. Wyatt currently work in the Distributed Autonomous Systems Laboratory (DASLab) at the University of Illinois at Urbana-Champaign.
Analysis of Musculoskeletal Serpentine Gait towards Rational Design of Soft Snake Robot
February 23, 11:30-11:50, CSL B02
Mimicking the biological serpentine gaits, snake robots may be capable of adaptive terrestrial locomotive strategies, overcoming rough surfaces and inspecting in narrow spaces. While a soft snake robot replicates the body compliance of the real creature, enabling smoother and potentially more efficient propulsive gaits. Biomimetic design of soft snake robots requires the modeling of soft and deformative slender bodies, making decisions on the muscular arrangement and activation pattern. To approach the optimal design, a numerical modeling method is created based on Cosserat rod theory, which is leveraged to capture the dynamics of all the biological components (spine, muscles, tendons) with individual soft filaments. By developing interconnection schemes, filaments representing different elements are assembled into a musculoskeletal snake that reacts to its muscular activity. Coupled with an optimizer (CMA-ES), our method generates the optimal design of a soft snake robot as well as its muscular activation pattern for fastest forward velocity under the selection of different number of muscle groups considering practical hardwire capacities.
Xiaotian Zhang is a PhD student in MechSE department working under Prof. Gazzola. He obtain BS from Zhejiang University in 2015, and MS from UIUC in 2017. His current research focus is on numerical simulation of soft slender bodies, targeting applications in computational soft robotics and biomimetics.