A feedforward-feedback control framework for disturbance rejection in multi-loop PI controlled servo actuators under...

This paper investigates disturbance rejection for multi-loop PI controlled electric servo actuators with external motion applied at one end and accurate displacement tracking required at the other. Two sensing scenarios are addressed: (i) measurable disturbance displacement and (ii) measurable disturbance acceleration only. Feedforward-feedback controllers are developed based on closed-loop error models to achieve disturbance rejection and compensate for transient-performance degradation. A neural-network-based disturbance velocity estimation method is introduced to improve estimation accuracy under noise and unmodeled dynamics. Experimental results show that, under a 1 Hz, 0.02 m sinusoidal disturbance, the proposed controllers achieve 95.8% displacement suppression and 95.6% acceleration suppression in Scenario (i), and 72.4% and 85.1%, respectively, in Scenario (ii). Swept-frequency, non-periodic disturbance, and load-variation experiments further verify the robustness of the proposed methods.