Prof Koshy George

Department of Electrical and Electronics Engineering

  1. Learning and Adaptation
  2. Complex and Intelligent Systems



The National Institute of Engineering, Mysore


Indian Institute of Technology, Madras


Indian Institute of Science
Ph. D.


  • 2006-2020 | Professor | PES Institute of Technology/PES University, Bangalore, India.
  • 2001-2005 | Assistant Professor | M. S. Ramaiah Institute of Technology, Bangalore, India.
  • 2000-2001 | Post-doctoral Associate | Yale University, New Haven, CT, USA.
  • 1998-1999 | Post-doctoral Researcher | Delft University of Technology, Delft, The Netherlands.
  • 1997-1998 | Post-doctoral Fellow | Seoul National University, Seoul, S. Korea.
  • 1996-97 | Project Associate | Indian Institute of Science, Bangalore, India.
  • 1990 | Engineer Trainee (R&D) | Tata Consulting Engineers, Bangalore, India.

Research Interest

  • All systems that surround us today – economical, political, ecological or social – consist of diverse, interconnected and interdependent entities which adapt to their environments. The overall behaviour of such systems is often unpredictable as a system of systems is inherently complex, nonlinear and time-varying. Complex systems have feedback loops with strongly interdependent variables possessing multiple equilibrium points and are possibly rather sensitive to initial conditions with small changes in the initial conditions likely precipitating large changes. Most natural phenomena are nonlinear as well as time-varying in nature. Consequently, the interconnection of simple dynamical systems can lead to seemingly inexplicable complex overall behaviour with the causes and effects not obviously related. Aristotle summed up well: “The whole is greater than the sum of its parts.” Complex intelligent systems are about indirect effects, and the principal thrust of my research in a broad sense is into the development of tools for analysis and design of a system of dynamical systems that perceive, reason, learn and adapt intelligently. The specific research areas are as follows:
  • 1. Networked Adaptive Systems: Cyber-physical systems (CPS) are engineered systems that are built from, and depend upon, the seamless integration of computational algorithms and physical components. It is expected that the advances in CPS will enable capability and adaptability that exceeds the embedded systems of today, and will transform the manner in which people interact with engineered systems. In this problem area, the focus is on the fact that CPS are feedback systems with networked sensing and/or actuation.

    Traditional control theory assumes continuous or discrete-time signals, where the controller continually or periodically observes the physical subsystem, and continually or periodically provides actuation to the plant. CPS systems are closed-loop or feedback systems, where typically sensors make measurements of physical processes, the measurements are processed in the cyber subsystems, which then drive actuators that affect the physical processes. The control strategies implemented in the cyber subsystems need to be adaptive (responding to changing conditions and uncertainties in the physical system and environments) and predictive (anticipating such changes).

    The principal focus of the proposed research problem is on networked adaptive systems that deal with the stability and performance of classes of systems with uncertainties and nonlinearities, when information between the compensator and the plant is passed through a wireless network, with the objective to mitigate the effects of congestion in communication paths (e.g., packet delays and packet dropout). Due to the presence of uncertainties and nonlinearities in the physical system, adaptive solutions are required. Preliminary work on networked adaptive systems with all systems assumed to be linear, time-invariant and discrete-time show good promise. There is sufficient scope to extend this to continuous-time or sampled-data systems, systems in the presence of disturbances and noise, systems with modelling uncertainties, systems that have time-varying parameters as well as nonlinear systems.

  • 2. Multiple Models: The multiple-model idealisation is the practice of building multiple models each of which makes distinct claims about the nature and causal structure giving rise to a phenomenon. Clearly, such an approach does not expect a single best model to be generated. Accordingly, this best suits with sciences that deal with complex phenomena. For example, multiple models of predation are constructed in ecology each of which contains different idealising assumptions, approximations, and simplifications. Chemists maintain both molecular and valence bond models of chemical bonding which make different incompatible assumptions. The United States National Weather Service uses three complex models of circulation patterns to provide high fidelity predictions of the weather. In the control context, multiple Kalman filters were used to improve the accuracy of state estimates. Some of the earlier applications of multiple models are tracking manoeuvring targets, control of mean arterial pressure and cardiac output and failure detection and identification.

    There appears two different ways in which multiple models are used. In one approach, the best amongst a set of models are chosen, and in the other, all models contribute in decision-making. The focus of this area of research is to delve into the details of the two approaches, to sort out the mathematical foundations of these approaches, or the lack of any, and put together the applications that required the use of multiple models for improved behaviour.

    A preliminary question that is to be answered is the need for a model itself. Data-driven learning systems are a current topic of research worldwide which focuses on techniques that are essentially model-free. These notions are to be crystallised from a mathematical perspective, and the manner in which such methods can dovetail into the multiple model methodology is to be looked into.

  • 3. Adaptive Control and Identification of Time-varying Systems: It is widely known that demands of advancing technology are the prime movers of new theoretical advances. In various areas like medical, neuroscience, marketing, and security surveillance, there are problems in which decisions are to be made quickly in the presence of uncertainties, especially where these uncertainties vary with respect to time. The time variations can be categorised as slow, fast or rapid, as well as periodic or random.

    In particular, there is tremendous interest in the adaptive identification and control of systems with rapidly varying parameters. These problems arise in diverse forms and disciplines including finance, sociology, and engineering. These problems are in general mathematically intractable. Adaptive methods available for identifying LTI plants are generally inadequate to deal with such time-varying problems. Preliminary work indicates that the multiple-model methodology can deal with different kinds of time-variations. In this area of research, these notions are to be placed within a mathematical framework.

Awards & Fellowships

  • 1987 | Dr Sir M. Visweshwaraya Birth Centenary Commemoration Celebration Prize, The Bowan Memorial Prize, Dr R. D. Begamudre Alias Rakesh Das Prize | University of Mysore.
  • 1987-1990 | MHRD (GATE) Scholarship | Indian Institute of Technology Madras.
  • 1990-1996 | Ph.D. Fellowship | Indian Institute of Science, Bangalore.


  • Fellow, The Institution of Electronics and Telecommunication Engineers (IETE)
  • Senior Member, Institute of Electrical and Electronics Engineers (IEEE)
  • Member, Institute of Engineering and Technology (IET)
  • Member, International Neural Network Society (INNS)
  • Member, Automatic Control & Dynamic Optimization Society (ACDOS), Indian National Member Organization of International Federation of Automatic Control (IFAC)
  • Life Member, Indian Society of Technical Education (ISTE)

List of Publications

  • Approximate optimal distributed control of nonlinear interconnected systems using event-triggered nonzero-sum games – V. Narayanan, A. Sahoo, S. Jagannathan and K. GeorgeIEEE Transactions on Neural Networks and Learning Systems, 30(5):1512-1522 (2019).
  • A multiple model approach to time-series prediction using an online sequential learning algorithm – K. George and P. Mutalik – IEEE Transactions on Systems, Man, and Cybernetics: Systems, 49(5): 976-990 (2019).
  • Adaptive control of a class of nonlinear time-varying systems with multiple models – K. George and K. Subramanian, Control Theory and Technology, 14(4):261-272, (2016).
  • Introducing robustness in model predictive control with multiple models and switching – S. Prabhu and K. George, Control Theory and Technology, 12(3):217-236 (2014).
  • A multiple-model approach to robust performance – K. George International Journal of Aerospace Innovations, 4(1&2):43-56 (2010).
  • Adaptive Control of Time-varying Systems Using Multiple Models – K. S. Narendra, O.A. Driollet, M. Feiler and K. GeorgeInternational Journal of Adaptive Control and Signal Processing, 17(2):87-102 (2003).
  • Generative Adversarial Networks for Histopathology Staining – A. Ganesh and K. GeorgeGenerative Adversarial Networks For Image-To-Image Translation, edited by A. Nayyar, Solanki and M. Naved, Elsevier (2021). (To Appear)
  • Adaptive Control Using Multiple Models: A Methodology – K.S. Narendra, O.A. Driollet and K. GeorgeModeling, Control and Optimization of Complex Systems: In Honor of Professor Yu-Chi Ho, edited by Weibo Gong and Leyuan Shi, Kluwer Academic Publishers, Boston, USA (2002).
  • A hierarchical approach for multi-class galaxy classification – M. Singhal, S.V. Hegde, R. Makam and K. GeorgeThe 17th IEEE India International Conference (INDICON) New Delhi, India (2020).
  • Automated segmentation of overlapping cells in cervical cytology images using deep learning – A. Umadi, K. Nagarajan, J.B. Venkatesha, A. Ganesh and K. GeorgeThe 17th IEEE India International Conference (INDICON), New Delhi, India (2020).
  • A hardware proof of concept of networked adaptive systems – S.B. Srinivasa, R. Makam and K. GeorgeProceedings of the 6th International Conference on Control, Automation and Robotics (ICCAR), Singapore (2020).
  • A two-step adaptive strategy for simple time-varying systems – K. Subramanian and K. GeorgeIEEE Region 10 Conference (TENCON), Osaka, Japan (2020).
  • Multiple Models: A Survey – K. George and K.S. Narendra – The 19th Yale Workshop on Adaptive and Learning Systems, Yale University, New Haven, CT, USA (2019).
  • Multiple models for decentralised adaptive control – R. Makam and K. GeorgeThe 15th IEEE International Conference on Control & Automation (ICCA), Edinburgh, Scotland (2019).
  • Staining of unstained histology using style transfer with colour-based segmentation – A. Ganesh, N.R. Vasanth, A.S. Ramaswamy and K. GeorgeThe IEEE Region 10 Conference (TENCON), Kochi, India (2019).
  • Meta-cognitive neural networks for adaptive identification and control – T.B. Sriram, M.S. Kashi, R. Ugarakhod, and K. GeorgeThe IEEE Region 10 Conference (TENCON), Kochi, India (2019).
  • Monopoly using reinforcement learning – E. Arun, H. Rajesh, D. Chakrabarti, H. Cherala and K. GeorgeThe IEEE Region 10 Conference (TENCON), Kochi, India (2019).
  • Improving Kannada OCR using a stroke-based approach – E. Arun, J. Vinith, C. Pattar and K. GeorgeThe IEEE Region 10 Conference (TENCON), Kochi, India (2019).
  • An inversion-based neuro-controller for robot manipulators – R. Ugarakhod, S. Abraham and K. George – The 6th India Control Conference (ICC), Hyderabad, India, (2019).
  • Generating playlists on the basis of emotion – G. Subramanian, J. Verma, N. Chandrasekhar, K.C. Narendra and K. GeorgeThe IEEE Symposium Series on Computational Intelligence (SSCI), Bangalore, India (2018).
  • On the use of dropouts in neural networks for system identification and control – S.M. Yadav and K. GeorgeThe IEEE Symposium Series on Computational Intelligence (SSCI), Bangalore, India (2018).
  • A statistical evaluation of vector-space models for text categorization – Y. Vijay, A. Sengupta and K. George The IEEE Symposium Series on Computational Intelligence (SSCI), Bangalore, India (2018).
  • Stability analysis of deterministic discrete-time adaptive systems with second level adaptation – R. Makam, S. Ramaiah and K. GeorgeThe International Conference on Signals and Systems (ICSigSys), Bali, Indonesia (2018).
  • Comparison of reinforcement learning algorithms applied to the cart-pole problem – S. Nagendra, N. Podila, R. Ugarakhod and K. GeorgeThe International Conference on Advances in Computing, Communications and Informatics (ICACCI), Manipal, India (2017).
  • Optimal detection of a particle source using cubical and spherical directional detector arrays – S. Srikanth, N.S. Chaya, S. Gurugopinath and K. GeorgeThe 2017 American Control Conference (ACC), Seattle, WA, USA (2017).
  • Adaptive control of a class of networked linear time-varying systems – R. Makam and K. GeorgeThe 2nd India International Conference (ICC), Hyderabad, India (2016).
  • Tracking rapidly moving signal sources using an array of sensors – K. George and V. Ravi – The 17th Yale Workshop on Adaptive and Learning Systems, Yale University, New Haven, CT, USA (2015).
  • Mitigating the effect of packet delay in a class of networked nonlinear adaptive systems – K. George and R. Makam – The 16th Yale Workshop on Adaptive and Learning Systems, Yale University, New Haven, CT, USA (2013).
  • A robust neural network based pulse radar detection for weak signals – A. Padaki and K. GeorgeThe IEEE International Radar Conference (RADARCON), Washington DC, USA (2010).
  • Adaptive Control of Simple Nonlinear Systems Using Multiple Models – K.S. Narendra and K. GeorgeThe 2002 American Control Conference, Anchorage, AL, USA (2002).
  • A systematic and numerically efficient procedure for stable dynamic model inversion of LTI systems – K. George, M. Verhaegen and J.M.A. Scherpen – The 38th IEEE Conference on Decision and Control, Phoenix, AZ, USA (1999).

Contact Details