Hierarchical Control of Hydraulic Manipulators via Identifier–Critic Scheme and Robust Control

Abstract

High-precision motion control of multidegree-of-freedom (multi-DOF) hydraulic manipulators remains a significant challenge due to their highly coupled mechanical dynamics and strong nonlinearities. To address these difficulties, this article proposes a hierarchical control framework to effectively deal with the rigid-body dynamics and actuator-level nonlinearities of hydraulic manipulators. In the upper level controller, an identifier–critic neural network is employed to estimate the manipulator dynamics online under an interval excitation condition. Based on the identified model, an approximate optimal position control policy is generated by minimizing a predefined performance cost function, enabling adaptive treatment of dynamic coupling and modeling uncertainties. At the lower level, a model-based robust nonlinear control law is designed for the hydraulic actuators to ensure exponentially convergent force tracking, thereby mitigating the effects of nonlinearities, disturbances, and unmodeled dynamics. The effectiveness of the proposed hierarchical control framework is validated through comparative experiments on a 6-DOF hydraulic manipulator. Experimental results demonstrate superior tracking accuracy and transient response compared with baseline methods, highlighting the potential of the proposed approach for high-precision motion control of complex hydraulic manipulators.