Improved Iterative Learning Control of Hydraulic Manipulators: An Inner–Outer Loop Framework

Abstract

Modeling hydraulic manipulators is inherently challenging, which significantly complicates controller design. Although iterative learning control leverages its data-driven capability to handle difficult-to-model systems, the strong nonlinearities in hydraulic manipulators limit its ability to achieve high-accuracy control. Notably, while formulating the mechanical dynamics model for hydraulic manipulators is impractical, hydraulic dynamic information remains accessible. To this end, we propose an inner–outer loop control framework to enhance the control performance of iterative learning control for hydraulic manipulators. In this framework, the outer loop handles position control of robotic manipulators using data-driven iterative learning control, while the inner loop focuses on the force tracking of the hydraulic actuator through a model-based nonlinear robust controller, ensuring exponential asymptotic tracking. The proposed method is validated through comparative simulations and experiments on a six-degree-of-freedom hydraulic manipulator. The results demonstrate that the inner–outer loop framework improves iterative learning control performance and offers a new approach to controller design for complex systems.