Adaptive frequency sampling based on local rational modelling

RF and mm-wave systems inherently exhibit a frequency-dependent dynamic behaviour (S-parameters). To characterize this behaviour, extensive electromagnetic simulations or measurements need to be performed. Unfortunately, these simulations and measurements take a lot of time, which ranges from minutes to hours. To save valuable characterization time, engineers need to carefully select at which frequencies they want to simulate/measure the dynamic behaviour. Adaptive frequency sampling techniques have been introduced for this purpose. They modify in real time at which frequencies the characterization should be carried out next based on what has been learned from previous frequency samples. The purpose is to improve the selection of frequencies and make it as time efficient as possible.

In recent years, we developed local rational modelling techniques at the department ELEC. As the name suggests, these techniques model the dynamic behaviour in a local frequency window with a rational model (see figure). The aim of this thesis is to develop a new adaptive frequency sampling scheme that uses the local rational models as information criterion. During this thesis you will

  • Study existing adaptive frequency sampling techniques.
  • Make the local rational modelling techniques compatible with adaptive sampling schemes.
  • Apply the developed technique to the characterization of RF and mm-wave systems, both in simulation as with actual measurements.

Interested or curious to know more? Feel free to contact Dries Peumans



Francesco Ferranti (B-PHOT, VUB)
Back to top