On real-time solutions for data-driven design, signal processing and control of noisy nonlinear systems KU Leuven
Mathematical models of dynamical processes provide the critical link between real-life applications and techniques for prediction, monitoring, and control. Models are obtained from observed data via system identification methods. These methods, however, do not take into account the subsequent use of the model for design. The issue is currently addressed by trial-and-error human interaction. A rigorous high-gain approach, investigated in this ...