摘要
Angular acceleration estimation and its application in acceleration feedback control are investigated experimentally in the paper. In combination of Newton Predictor (NP) with Kalman Filter (KF), a new predictive estimator for angular acceleration, called Newton Predictor Enhanced Kalman Filter (NPEKF) is proposed. This estimator provides a wide bandwidth and a small phase lag of the estimated acceleration while attenuating noises. Based on the estimated acceleration an acceleration feedback control (AFC) is presented for multiple degree-of-freedom (DOF) mechatronic system. The design of AFC is specified in terms of its stability and ability in suppressing dynamic disturbances. Experiments are conducted on a 2-DOF direct-drive manipulator. The frequency responses of the acceleration estimated by NP, KF and NPEKF are compared with those of the measured acceleration via linear accelerometer. The performance of AFC using the estimated acceleration is assessed against that using the measured acceleration. This study has shown that the proposed NPEKF estimator is able to supply the AFC with reliable required acceleration.