AbstracI-In this paper a new and completely linear algorithm is proposed for composite robust control of flexible joint robots. Maraver, the robust stability of the closed loop system in presence of structured and unstructured uncertainties is analyzed. To introduce the idea, flexible joint robot aith structured and unstructured uncertainties is modelled and converted into singular perturbation form. A robust linear control algorithm is proposed toor the slow dynamics and its robust stability conditions are derived using Thikhonov's theorem. Then the robust stability of the total system considering the porposed composite controller is analyzed, and sufftcient conditions for robust stability of system is obtained. Finally the elktiveness of the proposed controller is verified through simulations. It is shown that not only the tracking performance of the proposed controller is very suitable, but also the actuator effort is much smaller than previous result.
Abstract-RTAC1 benchmark problem considers a nonlinear fourth-order dynamical system involving the nonlinear interaction of a translational oscillator and an eccentric rotational proof mass. This problem has been posed to investigate the utility of a rotational proof mass actuator for stabilizing translational motion. In order to implement any of the model-based controllers proposed in the literature, the values of model parameters are required which are generally difficult to determine rigorously. In this paper, an approach to the leastsquares estimation of system parameters is discussed and practically applied to the benchmark problem. Next, in order to design an H ∞ controller, the nonlinear system is modelled as a perturbed linear system using an effective identification scheme. Experimental results confirm that this approach can effectively condense the whole nonlinearities, uncertainties, and disturbances within the system into a favorable perturbation block. Finally, an effective mixed-sensitivity problem is developed for the system to satisfy all performance requirements as well as robust stability despite actuator saturation.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.