Nonlinear Signal-Filtering Technique With Real-Time Gain Booster for Feedback System Applications

Abstract

This letter devises a novel signal-filtering technique subject to the feedback structure without the exact source signal dynamics information. The main idea of this result is comprised of two parts: first, the combination of the Luenberger and disturbance observer design techniques to improve the disturbance rejection capability in the feedback structure; second, the closed-form filtering gain-boosting mechanism updating the gain into the direction of the filtering error reduction. It is experimentally confirmed that the proposed filter contributes to lowering the current regulation performance degradation level from the filtering performance improvement using a 500-W permanent-magnet synchronous motor driven by the three-phase inverter.