End-effector Regulation of Robots with Elastic Elements by an Iterative Scheme

The regulation of motor variable positions in robots with elastic elements has been solved using PD linear controllers in the absence of gravity and with the addition of a model-based feedforward term when gravity is present. When the mass of the links is not known, an iterative learning scheme can be derived to obtain the same result for both joint elasticity and link flexibility. An extension to end-effector regulation with a similar two-stage and hence time-consuming scheme has been proposed. In this paper we show the feasibility of a new iterative one-stage scheme able to directly regulate the end-effector position in robots with joint elasticity and in robots whose distributed flexibility is limited for each link to the plane orthogonal to the associated motor axis. Experimental results are included to show the improved rate of convergence of the proposed scheme applied to a two-link flexible robot available in our laboratory.