Decoupled control of robot movement

Recently, methods have been proposed for the simultaneous decoupling of the controls of nonlinear dynamical systems and for the assignment of poles in the resulting decoupled linear systems. These methods have been used to obtain decoupled controls for the equations of motion of industrial robot arms and their performance has been gauged by simulation. It is shown here that such decoupling procedures do not guarantee stable controls and that their use, where the mathematical models of the controlled systems may be subject to uncertainty, should be attempted with caution.