Least Costly Space-Filling Experiment Design for the Identification of a Nonlinear System

TL;DR

This paper introduces a cost-effective space-filling input design method for nonlinear system identification that maintains model accuracy while reducing experimental costs.

Contribution

It proposes a novel approach to minimize experimentation costs in system identification while ensuring sufficient data quality for accurate modeling.

Findings

The method reduces experimental costs significantly.

It maintains model performance comparable to traditional designs.

Monte Carlo simulations validate the effectiveness of the approach.

Abstract

The quality of an estimated nonlinear model highly depends on the data quality that was used for the system identification. By using a Gaussian Process-based optimal input design approach, a so-called space-filling dataset can be generated in the feature space of the system model. The design method is applicable for a broad type of signals and models and also incorporates information measures through optimality criteria into the signal design. However, the resulting input design can be costly to apply to the real system. The goal of this paper is to propose a space-filling input design that can minimize the experimentation cost in terms of a user defined measure, while still guaranteeing a prescribed level of space-fillingness. Through a Monte Carlo simulation study we demonstrate that the proposed method can appropriately shape the excitation signal to significantly reduce the experimental cost while the identified model performance remains adequate.