Dynamic control of robots with joint elasticity

Reference is made to the problem of controlling the dynamic behavior of robots with rigid links but in presence of joint elasticity. It is shown that use of the more general class of dynamic nonlinear state-feedback allows solving both the feedback linearization and the input-output decoupling problems. A constructive procedure for the decoupling and linearizing feedback is given which is based on generalization system inversion and on the properties of the so-called zero-dynamics of the system. A case study of a planar two-link robot with elastic joints is included. The role of dynamic feedback for this class of robots is discussed