Optimized Excitation Signal Design Employing Receding Horizon Control

TL;DR

The proposed method is shown to effectively generate space-filling designs within the input space of a nonlinear dynamic process, thereby enabling sophisticated acquisition of information in previously unexplored operational areas.

Abstract

A novel excitation signal design strategy based on a receding horizon control inspired optimization is presented. The proposed method is shown to effectively generate space-filling designs within the input space of a nonlinear dynamic process, thereby enabling sophisticated acquisition of information in previously unexplored operational areas. Additionally, the strategy can intensify the exploitation of specific operational areas during information gathering, offering flexibility in meeting application-specific requirements.

Figures (1)

• Figure 1: Excitation signals and below the corresponding process input space distributions employing the example of a nonlinear first-order Hammerstein system. When calculating distances to the red-dotted points of $\underline{ {\Psi}}$, an increased weighting was applied.