Compensation Function Observer-Based Model-Free Adaptive Full-Order Sliding Mode Control for Hydraulic Manipulator with Payload Uncertainties

The trajectory tracking of hydraulic manipulator under complex operating condition is a challenging job due to its inherent strong nonlinearities and unknown external disturbances. In this paper, a compensation function observer-based model-free adaptive full-order sliding mode control for hydraulic manipulator is proposed in the presence of unknown payload and friction. To simplify controller design and overcome dependency on model, the ultra-local model (ULM) is utilized to approximate the complex hydraulic manipulator system in a short sliding time window. The compensation function observer is designed based on ULM to estimate the lumped uncertainties which is robust to the high-frequency disturbances and can ensure zero estimation errors. Moreover, full-order sliding mode surface is constructed to accelerate error convergence with a switching sliding gain adaption law based on boundary layer. The switching gain increase as the sliding motion moves away from boundary to improve the trajectory performance. In contrast, the gain will maintain a small value to alleviate the chattering problem. To validate the performance of the proposed control strategy, co-simulation of 3-DOF hydraulic manipulator are realized using SolidWorks, Simscape Multibody/Fluids, Matlab/Simulink in the presence of time-varying payload and joint friction. In addition, the performance of the proposed CFO-AFOSMC is demonstrated compared with TDE-iPD and TDE-SMC.