Parameter estimation of a nonlinear benchmark system

Abstract

This paper describes a possible procedure for the parameter estimation of a nonlinear benchmark problem called Silver box under the presumption that it has a structure of Duffing oscillator. The multiple-minima problem of optimisation is tackled with a combination of stochastic and deterministic optimisation methods. Simulation responses of the obtained model show a good match with the measurements taken from the electronic device, which is confirmed by validation of input/output response and residuals analysis.