Parameter Identification for Model-Based Control of Hydraulically Actuated Open-Chain Manipulators

In this paper, a nonlinear model-based controller with parameter identification is designed for a rigid open-chain manipulator arm actuated by servovalve-controlled hydraulic cylinders. The arising problem in adopting model-based controllers is how to acquire accurate estimates of system parameters, with limited available information about either the hydraulic actuator parameters or manipulator link inertial parameters. The objective of this study is to identify both the rigid-body parameters of the links and the hydraulic actuator parameters from collected cylinder chamber pressure and joint angle data, while no a priori knowledge of these parameters is available. Same physical plant models are used for control design as well as for parameter identification. Experimental results show that the proposed nonlinear model-based control scheme results in acceptable Cartesian position tracking performance in free-space motion when using the identified parameters.