4.7 Article

Compliant Control of Flexible Joint by Dual-Disturbance Observer and Predictive Feedforward

Journal

IEEE-ASME TRANSACTIONS ON MECHATRONICS
Volume -, Issue -, Pages -

Publisher

IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TMECH.2023.3272364

Keywords

Admittance control; disturbance observer (DOB); flexible joint; force estimation; impedance control; sensitivity; trajectory prediction

Ask authors/readers for more resources

This article proposes a novel dual-disturbance observer (DDOB) based on the encoder feedbacks from both the motor and link sides to achieve compliant control of a flexible joint. The DDOB estimates and compensates for friction and external torque separately without the need for a torque sensor. The results of simulations and real-time experiments demonstrate the practical appeal of this method.
The compliant control of a flexible joint relies on the accurate external torque information and effective internal disturbance compensation. To achieve this, most of the prior works use a built-in torque sensor and a lumped disturbance observer based on a single encoder. This increases the weight and cost of the system. In this article, a novel dual-disturbance observer (DDOB) based on the encoder feedbacks from both the motor and link sides is proposed so that the friction and external torque are estimated and compensated separately without the torque sensor. Thereby, a feedforward-feedback-DDOB composite scheme is formed for position control. The modified reference sensitivity of this scheme suggests that better tracking accuracy and disturbance rejection ability are achieved. In addition, the estimated external torque is used to alternate the reference trajectory with the given admittance model. To ensure that the feedforward control is realizable, the prediction of the alternated trajectory is done by the online fitting of a polynomial. The output turbulence caused by prediction errors is effectively suppressed by a single deadbeat compensator with the most ancient prediction in memory, while other predictions are weighted by the time-varying ratios. Simulations and real-time experiments are performed to demonstrate the practical appeal of the proposed method.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

4.7
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

No Data Available
No Data Available