Department of Adaptive Systems

In the first part of the talk author concentrates on methods of identification of nonlinear dynamic systems of "black box" type using integral-functional Volterra series approach. The key idea of the approach is to identify the transfer functions using special input perturbation functions. By these test signals it is possible to  "teach" the Volterra model. An identification problem is reduced to the solution of special Volterra integral equations of the first kind.

Existence and uniqueness of the solution is proven for these equations. Numeric aspects of the algorithm will be discussed too. The case of two input signals and one output signal will be discussed as important case of applications. Some details of applicability of this universal method in heat-and-power industry will be provided. 

The second part of the talk deals with adaptive nonlinear filter called empirical mode decomposition which is a key part of  Hilbert-Huang Transform.