Intelligent Systems and Robotics

Priorities, axes and perspectives in industrial systems technology and in particular in the vital area of nonlinear control and nonlinear estimation are related with: (1) Methods based on global linearization-based control (2) Methods based on asymptotic linearization-based control, and (3) model-free Lyapunov theory-based methods of control.

As far as approach (i) is concerned, that is methods based on global linearization these are methods for the transformation of the nonlinear dynamics of the system to equivalent linear state-space descriptions for which one can design controllers using state feedback and can also solve the associated state estimation (filtering) problem. One can classify here methods based on the theory of differentially flat systems and methods based on Lie algebra. These approaches avoid approximate modelling errors and arrive at controllers of elevated precision and robustness. In this area, research will be continued on a new nonlinear filtering method (based on differential flatness theory) which has been demonstrated to have excellent performance, in comparison to other nonlinear estimation (filtering) approaches, both in terms of accuracy and in terms of speed of computation. 

As far as approach (ii) is concerned that is methods based on asymptotic linearization, research will be carried out on robust and adaptive control through decomposition of the system’s dynamics into local linear models.  In this research area, the institute’s work will pursue solutions to the problem of nonlinear control with the use of local linear models (obtained at local equilibria). For such local linear models, feedback controllers of proven stability can be developed.  One can select the parameters of such local controllers in a manner that assures the robustness of the control loop to both external perturbations and to model parametric uncertainty. These controllers achieve asymptotically (that is as time advances) the compensation of the system’s nonlinear dynamics and the stabilization of the closed control loops. In this area, joint research will focus on a new method of robust control, being developed by me during the last years. The method is based on local approximate linearization of the system’s dynamics and solves the nonlinear H-infinity control problem. 

As far as approach (iii) is concerned, that is methods of nonlinear control relying on Lyapunov stability theory I have successfully handled problems of minimization of Lyapunov functions so as to assure the asymptotic stability of the control loop of nonlinear control systems. For the development of Lyapunov type controllers one can either exploit a model about the system’s dynamics or can proceed in a model-free manner, as in the case of indirect adaptive control. In the latter approach, the system’s dynamics is taken to be completely unknown and can be approximated by adaptive algorithms, which are suitably designed so as to assure the stabilization and robustness of the control loop. Research in this area is anticipated to arrive at new and significant results.

Application areas

  • Electric Power Systems. Research on the following technologies for renewable energy systems: (i) intelligent control of power generators (ii) intelligent control of the power electronics that connect renewable power generation units to the grid (iii) synchronization and stability of distributed power generation units (iv) fault diagnosis for power generators and power electronics.
  • Industrial Robotic Systems. Research on nonlinear control and estimation for Industrial robotics: model-based and model-free control approaches for multi-DOF robotic manipulators, underactuated robots, closed-chain robotic mechanisms and flexible-link robots. Model-based and model-free control for mechatronic systems (such as electric actuators).
  • Intelligent Transportation Systems. Development of nonlinear control and filtering methods for autonomous robots and vehicles of various types, such as Unmanned Ground Vehicles (UGVs), Unmanned Aerial Vehicles (UAVs, Unmanned Surface Vessels (USV,) or Autonomous Underwater Vessels (AUV). Development of nonlinear control and estimation methods for electric motors and combustion engines.
  • Biosystems. Applications of nonlinear control and filtering to biosystems aiming at modifying their dynamics and at achieving their functioning according to specific performance objectives. Automatic control methods in protein and hormonal synthesis or in the infusion of cardiovascular, chemotherapy and anaesthesia drugs. Nonlinear estimation techniques for modelling and detection of degeneration in biological neuronal systems.