A Complementary Estimation Scheme of Angular Accelerations for Robot Manipulators under Slow Sampling Rates

An accurate estimation of angular acceleration signals for robot manipulators is necessary for the improved control performance. In this paper, a complementary estimation method for angular acceleration of robot manipulators under slow sampling rates is proposed. Based on the data-driven system model, a state observer is constructed to estimate acceleration signals, which require fast sampling rates to improve the performance. Since a linear model-based state observer with slow sampling time undergoes low-frequency sensitivity, a complementary scheme combining the finite difference approach and the state observer approach together is proposed. By fusing them, we could obtain an improved estimation of angular acceleration signals with low-frequency robustness.