In this paper, impedance learning is investigated for robots interacting with unknown environments. A two-loop control framework is employed and adaptive control is developed for the inner-loop position control. The environments are described as time-varying systems with unknown parameters in the state-space form. The gradient-following and betterment schemes are employed to obtain a desired impedance model, subject to unknown environments. The desired interaction performance is achieved in the sense that a defined cost function is minimized. Simulation and experiment studies are carried out to verify the validity of the proposed method.
Mendeley saves you time finding and organizing research
Choose a citation style from the tabs below