Predictive Identification of Electromagnetic Micro-mirrors

This paper presents an identification approach for electromagnetic micro-mirrors using data-driven model with hysteresis. Usually, linear models are applied to modeling of electromagnetic micro-mirrors. However, from the experimental results, it can find that hysteresis phenomenon exists in the electromagnetic micro-mirror. Thus, it is not proper for using the linear model for modeling of electromagnetic micro-mirror. In this paper, a data-driven model with hysteresis is proposed for modeling the performance of electromagnetic micro-mirrors. In this model, a linear dynamic submodel is employed to model the deflection mechanism, whilst a dynamic hysteresis submodel is used to describe the hysteresis phenomenon in the electromagnetic actuating mechanism of the micro-mirror. One of the challenges of identification for electromagnetic micro-mirror is the electromagnetic torque cannot be measured directly. For handling this problem, a so-called predictive identification algorithm is applied to parameter estimation of the model. The effect of random noise on the identification is also considered. Then, the paper also presents the experimental results of modeling an electromagnetic micro-mirror.