• Stability and robustness of neral network model predictive control
    Research concerns stability conditions of nonlinear model predictive control based on GPC using neural network input-output models. In order to obtain a monotonically decreased cost function in time terminal constraints have been used.
    See the presentation given at SysTol 2013, Nice, France.
    The second issue is the analysis of the predictive control in the presence of noise, disturbances and model uncertainty. To deal with the robustness of the control scheme, a Model Error Modeling technique has been used. The derived uncertainty region has been then applied to redefine the open-loop nonlinear optimization problem.
    See the presentation given at MSC 2014, Antibes, France.
  • Fault tolerant control using state-space neural models
    Research concerns fault compensation in an existing control system equipped with the PID controller. Using a neural state-space model as well as a neural observer, one can determine an additional control signal. Then, when a fault occurs in the system, the control system is able to determine an estimate of an unknown fault function, and fault accommodation is carried out. Thus, the control system tries to perform as close as possible to the required normal operating conditions.
    See the presentation given at DPS 2011, Zamość, Poland.
  • Non-linear iterative learning control Research concerns designing non-linear iterztive learning control by means of artificial neural networks. A dynamic neural network plays a role of a model of a system while simple static feeedforward net is used as the learning controller. Convergence analysis of the proposed control schemes is also investigated.
    See the presentation given at ECC 2018, Limassol, Cyprus.
    See the presentation given at ACC` 2019, Philadelphia, USA.
    See the presentation given at IJCNN 2021, Shenzen, PR China.
  • Seizure classification Research concerns the detection of epileptic seizures using deep Long Short-Term Memory neural networks.
    See the presentation given at DPS 2020, Zielona Góra, Poland.
  • Neural modeling of nonlinear distributed parameter systems
    The research concerned the use of a set of dynamic neural networks to build a model of distributes parameter system. The idea was to divide the entire spatial space into smaller partitions and for each partition to built a particular neural model. Reconstruction of the object's response over the entire space was based on interpolation of responses of individual neural models.
    See the poster - IJCNN 2022, Padova, Italy
    Another approach was to use the Echo State Network to realize the so-called spatial reservoir.
    See the presentation - IFAC Congress 2023, Yokohama, Japan
Recent publications
  • Fault-tolerant design of non-linear iterative learning control using neural networks. Patan K., Patan M., Engineering Applications of Artificial Intelligence, Vol. 124, paper 106501, available online, 2023.
    share link to the full version
  • Neural-network-based iterative learning control of nonlinear systems. Patan K., Patan M, ISA Transactions, Vol. 98, pp. 445-453, DOI: 10.1016/j.isatra.2019.08.044, 2020.
  • Two stage neural network modelling for robust model predictive control. Patan K., ISA Transactions, Vol. 71, pp. 56-65, DOI: 10.1016/j.isatra.2017.10.011, 2018.
Most cited publications
  • Neural network based model predictive control: Fault tolerance and stability. Patan K., IEEE Transactions on Control System Technology, Vol.23, No. 3, pp 1147-1155, DOI: 10.1109/TCST.2014.2354981, 2015.
  • Soft computing approaches to fault diagnosis for dynamic systems. Calado J.M.F., Korbicz J., Patan K., Sa da Costa, J. M. G., Patton R. European Journal of Control, Vol. 7, No. 7-8, pp. 248-286, DOI: 10.3166/ejc.7.248-286, 2002.
  • Identification of neural dynamic models for fault detection and isolation: the case of a real sugar evaporation process. Patan K., Parisini T., Journal of Process Control, Vol. 15, No. 1, pp. 67-79, DOI: 10.1016/j.jprocont.2004.04.001, 2005.
  • Towards robustness in neural network based fault diagnosis. Patan K., Witczak M. Korbicz J., International Journal of Applied Mathematics and Computer Science, Vol. 18, No. 4, pp. 443-454, DOI: 10.2478/v10006-008-0039-2 , 2008.
  • Stability analysis and the stabilization of a class of discrete-time dynamic neural networks. Patan K., IEEE Transactions on Neural Networks, Vol. 18, No. 3, pp. 660-673, DOI: 10.1109/TNN.2007.891199 , 2007.
Full publication list
© Krzysztof Patan