Robot manipulator control by acceleration feedback

Luo and Saridis have shown that acceleration feedback is optimal with feedforward compensation of the non-linear terms of the equations of motion. In this note, the effect of these non-linearities is explored by use of a Lyapunov stability analysis under LQ assumptions. As a result of this analysis, it is shown that a bounded stable region exists. This region defines a steady state error and maximum operating range. It is shown that this region can be made arbitrarily large, reducing the steady state error and increasing the maximum operating range.