Flexible-Joint(FJ)robots have attracted researchers'attention in these years for their good compliance and the safety.But as a matter of fact,it is hard to design a suitable controller with stiffness adaptation for FJ robots due to the model uncertainties and the nonlinear systems when using elastic components.In this paper,an iterative learning control method for flexible-joint robot with model uncertainties in both robot dynamics and actuator dynamics is proposed,which involves the use of mechanical elastic elements with adjustable stiffness.Adjusting the stiffness of the joints by using the positions of the roller installed inside of the actuator,the joints generate fixed periodic motions and reduced the tracking error.We show the theory of the variable stiffness and the adaptive law,which involve the iterative learning controller.Simulation results also demonstrate that we achieve a perfect result while generating the desired motions.Index Terms:iterative learning;stiffness adaptation;flexible-joint robot;motion tracking.
展开▼
