Professor Nesic's research interests are in the broad area of control engineering including its mathematical foundations (e.g. Lyapunov stability theory, singular perturbations, averaging) and its applications to various areas of engineering (e.g. automotive control, optical telecommunications) and science (e.g. neuroscience). More specifically, he has made significant contributions to the areas of nonlinear sampled-data systems, nonlinear networked control systems and extremum seeking control and he presented several keynote lectures on these topics at international conferences.
Underlying most of his work is the Lyapunov stability theory and its different extensions
to systems with disturbances within the framework of input-to-state stability (ISS). Also, he has done
extensive work on the averaging and singular perturbations methods for nonlinear systems with
disturbances. His PhD Thesis was on analysis of controllability of discrete-time polynomial (nonlinear)
systems. He also developed a novel framework for controller design for nonlinear sampled-data systems via their approximate
discrete-time models and applied these tools to a range of engineered systems including multi-robot systems, internal combustion engines, atomic force microscopes, Lithium-Ion batteries and Raman optical amplifiers. His more recent work is on hybrid and networked control systems, event-triggered control, privacy and security in cyber-physical systems and optimization-based control for general nonlinear systems, including value and policy iteration methods.
| Back | Curriculum Vitae | Research | Complete Publication List | Google Scholar | Teaching | Contact | Department Page |
When stabilization and optimization meet: a codesign approach, D. Nesic, R. Postoyan and L. Busoniu, ARC Discovery Project, 2021-2023.
Enabling the Internet of Things (IoT): structured networked control systems, D. Nesic, Y. Tan, R. Postoyan, M. Heemels and A.R. Teel, ARC Discovery Project, 2020-2022.
Observer design for singularly perturbed systems, D. Nesic and C. Manzie, ARC Discovery Project, 2017-2021.
Modeling, analysis and design of secure networked control systems, D. Nesic, I. Shames, R. Postoyan and A.R. Teel, ARC Discovery Project, 2016-2019.
Defense Cooperative Research Centre on Trusted Autonomous Systems, C. Manzie et al, ARC, 2019-2021.
C. Manzie, I. Shames, D. Nesic and H. Nakada, Practical model-based control for low emission and low-cost diesel engines, ARC Linkage Project, 2016-2019.
M. Cook, D. Grayden, H. McDermot and D. Nesic, Monitoring Cortical Excitability using a Probing Stimulus for Epileptic Seizure Anticipation, , NMHRC Project Grant, 1048360, 2013-2015.
D. Nesic, Y. Tan and P.M. Dower, Extremum seeking control: a systematic design framework, ARC Discovery Project, DP120101144, 2012-2014.
D. Nesic, Analysis and Design of Networked Control Systems, ARC Discovery Project, DP1094326, 2010-2015.
D. Nesic, Networked Control Systems: Harnessing An Emerging Technology, ARC Future Fellowship, FT0990727, 2009-2013.
D. Grayden, D. Freestone, M. Cook and D. Nesic, Optimisation Of Signal Processing And Electrical Stimulation Algorithms For The Abatement Of Epileptic Seizures, ARC Linkage Project, LP100200571, 2010-2012.
D. Nesic, Y. Tan, C. Manzie and I.M.Y. Mareels, Extremum seeking control: analysis, design and applications, ARC Discovery Project, 2009-2011.
T. L. Song, D. Musicki and D. Nesic, Feedback Enhanced Sensor Signal Processing for Robust Target Tracking Using Underwater Sensors, ADD-09-70-01-03, 2009-2011.
M. Brier, H. Watson, C. Manzie, W. Ducker and D. Nesic, Efficient and practical hydrogen fuelled vehicle technologies, Department of Innovation, Industry and Regional Development, Victorian Government, Energy Technology Innovation Strategy (ETIS), SU504122, 2007-2010.
D. Nesic, Analysis, design and applications of hybrid systems - a new perspective on complex engineering problems, Melbourne Research Grants Scheme, 600764, 2008.
D. Nesic and Y.Tan, Finite-dimensional sampled-data control of nonlinear spatially distributed parameter systems, ARC Discovery Project with an Australian Postdoctoral Fellowship (APD), 2006-2008.
D. Nesic and L. Zaccarian, Analysis and design of control systems with saturation and time delay, ARC International Linkage Project, 2005-2007.
D. Nesic, France, Stability of hybrid systems in application to network controlled systems, EGIDE Scholarship, 2005.
D.Nesic and 30 other researchers, Seed funding for Network on Control, Dynamics and Systems, ARC Special Initiatives, 2004.
D. Nesic, New research directions in sampled-data nonlinear systems, ARC Discovery Project with an Australian Professorial Fellowship (APF), 2004-2008.
D. Nesic, Redesign of continuous time nonlinear controllers for digital implementation, Alexander von Humboldt Fellowship, 2003-2004.
D. Nesic, I. Mareels and P.Dower, Nonlinear systems with disturbances: analysis, controller design and tradeoffs, ARC Discovery Project, 2003-2005.
D. Nesic and P. Dower, Tools for nonlinear continuous-time control systems with disturbances, ARC Discovery Project, 2002.
D. Nesic, Sampled-data nonlinear control systems and numerical methods, ARC Large Project, 2001-2003
D. Nesic, Analysis of properties of time-varying systems with disturbances via averaging, ARC Small Project, 2000.
I. Mareels and D. Nesic, Control of nonlinear systems described by polynomial equations, ARC Large Project, 1998-2001.
D. Nesic, Linkoping University, Sweden, Travel Grant, 1997.
D. Nesic, Catholic University, Louvain la Neuve, Belgium, Travel Grant, 1997.
D. Nesic, Australian National University, Travel Grant, 1995-1996.
When stabilization and optimization meet: a codesign approach
Discovery Project funded by the Australian Research Council
Grant number: DP210102600 | Funding period: 2021-2023
Investigators: D. Nesic (CI), R. Postoyan (PI) and L. Busoniu (PI)
Research Fellows: Mathieu Granzotto
PhD Students: Shuhao Qi
Collaborators: Jamal Daafouz
Summary:The next generation of engineered systems need to perform complex tasks with precision, and be robust, resilient and adaptive to their environment enabled by the confluence of control, optimization, learning and computation Understanding the interplay between robust stability and optimization is key to this endeavor. Many techniques, such as model predictive control and reinforcement learning, rely on an intricate interplay between an optimization-based control algorithm and an optimization routine used to calculate the control law. This project aims to develop a general design framework for stability, suboptimality and robustness of such algorithms, that can be used in range of novel applications, such as driverless cars and drones.
This project will enable the design of advanced control algorithms used in cyber-physical systems by understanding in depth the stabilizing, near-optimality and robustness features of optimization based control algorithms, such as model predictive control and approximate dynamic programming. The capabilities of these cyber-physical systems will be enhanced considerably by careful designs of their ``brain", which can learn about their environment, adapt to it and perform complex tasks with precision. Operating autonomously, the next generation of engineered systems will be essential for smart highways, driverless cars, swarms of drones, various types of robots and advanced manufacturing systems to name a few examples. It will impact transportation, environmental monitoring and defence, improving our quality of life in overpopulated cities, providing better use of our energy and water, reduce pollution and waste as well as maintaining a competitive edge in the global market.
Enabling the Internet of Things (IoT): structured networked control systems
Discovery Project funded by the Australian Research Council
Grant number: DP200101303 | Funding period: 2020-2022
Investigators: D. Nesic (CI), Y. Tan (CI), R. Postoyan (PI), M. Heemels (PI) and A.R. Teel (PI)
Research Fellows: Alejandro Maass
PhD Students: Weixuan Wang, Elena Petri
Collaborators: Daniele Astolfi
Summary:Networked control systems are an emerging technology that combines control, communication and computation to deliver solutions for a range of manufacturing, safety-critical infrastructure, such as transport, defence and other Industrial Interent of Things (IIoT) applications. The current analysis and design approaches often take a ``monolithic" view of the system, which render them inadequate for addressing many important IIoT applications. This proposal will exploit specific features and structure of the plant, the communication network and the distributed computation to provide an analysis and design methodology which will deliver significant advances in control and optimised performance of IIoT with benefits to the economy and society.
A range of manufacturing, safety-critical infrastructure, such as transport, and defence applications will increasingly rely on engineered solutions that exploit a fusion of control, communication and computation to control and optimise their performance. Further, companies globally are moving towards "digitalization of their business" and they want to use Internet of Things as a paradigm that drives their transformation. This project will deliver fundamental research on structured networked control systems that will enable significant advances in control, robustness and performance to deliver benefits to a range of novel applications and eventually make IoT/IIoT a reality. The potential benefits to the economy and society are enormous, including improvements in productivity and outputs in manufacturing and process industries, transportation and communication, energy efficiency, reduced pollution, regulation of our built environments and sophisticated devices for medical applications.
Observer design for singularly perturbed systems
Discovery Grant funded by the Australian Research Council
Grant number: DP170104102 | Funding period: 2017 - 2021
Investigators: D. Nesic (CI) and C. Manzie (CI)
Research Fellows: Mohammad Deghat and Saeed Ahmadizadeh
PhD Student: Luis Cuevas, PhD Thesis, University of Melbourne 2019, A general estimation framework for nonlinear singularly perturbed systems.
Collaborators: Andy Teel (UCSB), Romain Postoyan (University of Lorraine) and Zhiyong Sun (TU Eindhoven)
Summary:In many physical and engineered systems we are interested in knowing how certain variables change over time. The easiest solution is to measure them directly using different types of sensors, such as thermometers, barometers, accelerometers, dynamometers, and so on. However, it is often too expensive to measure all variables of interest or sometimes even impossible. In such situations, it is possible to estimate certain unmeasured variables by using other measurements and a model of the system under consideration. Algorithms that allow us to do this are referred to as estimators, filters or observers, depending on which literature you refer to. For instance, Kalman filter is a specific type of an estimator that provides optimal estimation under certain conditions. Observers are widely used in all areas of engineering, as well as physics, meteorology, neuroscience, finance, and so on. Observer design for systems that require nonlinear models is much harder than for linear models.
Figure 1: In this figure, the plant is a physical process of interest (e.g. a drone, car, robot, space shuttle), the actuator signals are denoted by u (control inputs), the measured variables are denoted as y (outputs) and the estimated variables are denoted as ^ x. Note that the observer is an algorithm (e.g. implemented on a computer) that uses the measured variables u and y to produce estimates ^ x of the unmeasured variables x.
This project concentrates on systems with nonlinear models that exhibit the so-called time-scale separation, or the so-called singularly perturbed systems. This further complicates observer design. These are systems where some physical variables in the system change very fast and others change very slowly. For instance, many examples of systems with time-scale separation arise in robotics, where mechanical variables (positions, velocities) change much more slowly than the electrical variables (currents, voltages). Actually, all engineered systems undergo ageing where some properties of the system slowly deteriorate over time as compared to other variables in the system; for instance, the state of the health of a battery in a drone changes slowly compared to currents and voltages that are used to drive the propellers.
Figure 2: Natural and engineered systems that exhibit a time scale separation: a) continuously stirred tank bioreactor, b) a single neuron, c) human brain, d) larvae preypredator system, e) DC-DC converter, f ) lithium-ion batteries, g) electrical motor, h) suspension system. (Taken from L. Cuevas PhD Thesis "A general estimation framework for nonlinear singularly perturbed systems")
This project contributed fundamental results on design of observers (i.e. estimators) for systems that exhibit time-scale separation. Prior to our work, very few results existed on a systematic observer design for general nonlinear singularly perturbed systems. Our results significantly pushed the state-of-the-art in this area and provided numerous novel results and observer design techniques for two time scale systems. Our results are stated as "frameworks" that cover a range of existing observer design techniques, such as high gain observers and circle criterion observers. Several observer design networks were developed for different classes of general nonlinear systems. In particular, a framework for observer design of slow variables for Tikhonov singularly perturbed systems was developed. A similar design framework was developed for design of observers for more general singularly perturbed systems that require averaging to obtain the slow reduced model. Several results on averaging and closeness of solutions for such systems was also developed as this is essential for proving convergence properties of the observers. Finally, a multi-observer approach was developed for estimation of slow and fast variables of parameterized systems with slowly varying parameters.
Our results have very wide applicability. Indeed, We demonstrated this in Luis Cuevas' PhD thesis by applying our results to flexible joint robots, car suspension system, vehicle dynamics and neural mass models that arise in modeling of the onset of epileptic seizures. Moreover, our results are expected to improve condition monitoring in mechanical, electrical and power systems. For example, our results can be applied for health monitoring of the state-of-health of electrical batteries. It has been hypothesised that slow changes in brain dynamics can be used to predict the onsets of epileptic seizures in some patients via EEG measurements and neural mass models and we are currently exploring this research direction.
As a part of this project, Luis Cuevas completed his PhD titled A general estimation framework for nonlinear singularly perturbedsystems in 2019. He is currently employed at Hit IQ as a research/data analyst. Also, Mohammad Deghat worked with us as a research fellow for two years; he is now an academic at the University of New South Wales. Saeed Ahmedizadeh also contributed to the project while he was working with us as a research fellow and currently he is working for Deloite. We have also collaborated with Prof. Andy Teel from the University of California in Santa Barbara, Dr Romain Postoyan from the University of Lorraine and Prof. Zhiyong Sun from the Australian National University.
Modeling, analysis and design of secure networked control systems
ARC Discovery Project (DP170104099), 2017-2019
D. Nesic (CI), I. Shames (CI), R. Postoyan (PI), A.R. Teel (PI)
Summary: Advances in computation and communication technology have lead to a new generation of Networked Control Systems (NCS) that have an enormous potential to control large-scale and complex distributed systems. Improved NCS technology will underpin our ability to optimise water and energy use, live in sustainable communities and create greater efficiencies in manufacturing and transport globally. These significant benefits can only be delivered through the development of novel NCS design methodologies essential to harnessing this emerging technology. We will address this challenging task by developing models, fundamental science and control engineering design techniques for NCS with several important physical networks.