Automatic Control and Systems team

The Automatic Control and Systems team works around two main axes: identification and control of systems.

Within the framework of the identification, the team is working on modelling, identification and diagnosis of nonlinear systems governed by partial derivative equations. The aim is to bring knowledge of the system which is to be analysed, controlled, observed or supervised. Numerous developments led in the LIAS allow the team to propose new algorithms, to make their use easier, to improve their convergence and to give access to the physical knowledge of the systems. The team also works in the field of system diagnosis and has developed a useful supervision methodology via parameter estimation, mainly applied to electric machines. Moreover, due to the development of an original fractional integration operator, the team is deeply involved in the research on modelling and identification of diffusive interfaces.

The second axis is the control of multi input multi output (MIMO) systems. A part of the work concerns the analysis, i.e. the research for numerically tractable conditions which permit to assess whether the system under study is stable or not, or whether it reaches some performance level (satisfactory transient behaviour, disturbance rejection, …) or not. Therefore the team is also interested in the control, i.e. the design of feedback control laws which allow the closed-loop system to attain the above-mentioned performances. The class of the systems which are studied in the LIAS is wide: It encompasses the linear and nonlinear systems, the uncertain systems, the fractional systems, the D systems, the time-delay systems, and so forth. Numerous practical applications are considered in the field of electrical engineering and renewable energy.

Scientific problems under study

  • Modelling and identification of: MIMO continuous-time systems, linear parameter-varying (LPV) systems and diffusive interfaces,
  • Closed-loop identification,
  • System diagnosis by parameter estimation and artificial intelligence,
  • Analysis and control of: fractional systems (non integer order models), nD systems (repetitive processes, PDE), descriptor, periodic and/or positive models, time delay systems,
  • Robust pole placement,
  • Active filtering.

Application domains

  • Transport (hybrid vehicles, flex fuel engines),
  • Telecommunications,
  • Management and distribution of energy (Smart Grids...),
  • Renewable energy (solar, energy, wind turbines, marine turbines...),
  • Machines électriques,
  • Power machines, electric motors,
  • Robotics (cable robot manipulators, haptic interfaces...),
  • Heat exchangers,
  • Waste water treatment,
  • Virtual sensors,
  • Control of fluid dynamics,
  • Harmonics filtering on the electricity network.

Set up methods, techniques and approches

  • System identification: output error, subspace algorithms, closed loop, continuous time neural networks, LPV and non linear model structure; initialisation of output error algorithms, convergence analysis, reinitialised partial moments, system identification using a priori information,
  • Modelling: electric motors in faults, magnetic circuits, PDE based systems (partial differential equation);
  • Simulation: finite element method, PDE based systems,
  • Control: LMI approach, S-procedure and other KYP-like techniques, complex parameter systems, quaternions, robust DU-stability, S-regularity, "Euler-Lagrange"-like approaches, eigenstructure assignment.