Robots control by adaptive gain smooth sliding observer-controller and parameter identification

Abstract

An adaptive gain, smooth sliding observer-controller are developed to control uncertain parameters, n-degree of freedom rigid robotic manipulators. Furthermore, an on-line, closed loop identification scheme, for time-varying parameters is proposed in order to obtain useful information despite loads, external disturbances and faults detection. In order to reduce the chattering, a smooth switching function (parameterised tangent hyperbolic function) is used instead of pure relay one, into the observer and the controller. The gains of the switching functions are adaptively updated, depending on the estimation error and tracking error, respectively. By using adaptive gains, the transient and tracking responses are improved. Simulation results with a two degree of freedom (DOF) robot manipulator are presented to show the interest of the approach.