Blind Identification for Controlled Limit-Cycle Systems Using Harmonic Balance and Output Oversampling

摘要

A new system identification technique is introduced for identifying controlled nonlinear feedback systems operating in a limit cycle. For this sake we use only one measurement variable in the loop for estimating both the linear dynamics in the feedforward loop and the static nonlinearity in the feedback path, and apply the harmonic balance principle together with the output oversampling technique. It shows that even though very limited information is available for the identification task, the linear part of the system is estimated in an acceptable level. But owing to the rolled-off nature of the output spectrum of the linear dynamics, the static nonlinearity cannot be accurately estimated without further information on it. The developed algorithm is demonstrated using a numerical example.