Robust Command-Filtered Control of Compliant Actuator-Driven Robotic Manipulators by Employing Beneficial Disturbances

Yang Zhang,Fuxin Du,Changwei Yin,Menghua Zhang,Rui Song,Yibin Li
DOI: https://doi.org/10.1109/wrrc62201.2024.10695964
2024-01-01
Abstract:A novel tracking control method is proposed for compliant actuator-driven robots in presence with parameter uncertainties, unknown nonlinear system dynamics, and external disturbances. The main feature of the proposed method is that beneficial parts of the lumped disturbances are retained to further improve tracking performance, rather than disturbances being fully seen as harmful elements to eliminate. Besides, its implementation requires only one system parameter, i.e., the motor inertia, which greatly eliminates the requirement for a priori knowledge of model parameters. Firstly, two extended state observers are designed to estimate unknown disturbances and unmeasurable system states. With the help of estimated information, disturbance effect indicators are constructed to evaluate whether the disturbance is beneficial or detrimental concerning current system motion. Then, a novel disturbance effect-triggered backstepping controller is proposed, in which the output tracking error of the system is guaranteed to converge into a small neighborhood of origin. In addition, it should be pointed out that the problem of “complexity explosion” is successfully avoided by introducing the filters into each step of the backstepping controller. Finally, the excellence of the recommended control method is validated by experiments.
What problem does this paper attempt to address?