Model-Based Variable-Structure Adaptive Control of Robot Manipulators

Model-based control enables the nonlinear, coupled and time-varying dynamics of multi-joint robot manipulator to be largely compensated. There are, however, inevitable uncertainties in the model and various adaptive control measures have been proposed to retain the inherent advantage of the model-based control. This paper deals with a recently developed scheme which singles out the uncertain parameters in the model and applies a switching control law acting on the parameter estimates understood to vary within given bounds. The particular features of the new method are examined by means of an experimental study.