Control System of Three Degree Freedom Parallel Manipulator

With parameter inaccuracies and bounded disturbances, a robust control scheme based on velocity observation is proposed for robot systems in which only position can be precisely measured. By the introduction of a compensation error in the design of the sliding mode observer that counteracts the effect of system deviations to the velocity reconstruction, the exponential observation error is converged to a bounded region. Using the Lyapunov direct method in designing the robust control laws based on the normal model and observed velocity, the closed-loop system satisfies the local uniform ultimate boundedness. The algorithm attends to tracking control of a limited freedom robot manipulator with smooth adaptive robust friction compensation. The new approach offers the ability to learn, to handle various uncertainties, to reject external disturbances, to ensure smooth control input, to improve transient performance, and to achieve global asymptotic stability