Single-loop versus multi-loop control strategies for electrically driven robots with elastic joint

A comprehensive comparison between single-loop and multi-loop control of electrically driven robots with elastic joint (EDREJ) is addressed in this paper. It should be emphasized that most of previous approaches presented for EDREJ utilize back-stepping-based control strategy. In order to have a satisfactory performance in these approaches, the internal signals should converge to their desired trajectories defined by the designer. Usually, these desired trajectories are known as fictitious control signals. In flexible-joint electrically driven robots, each joint is modeled by a 5th-order cascade differential equation. Thus, back-stepping-based approach seems complicated and time-consuming. Therefore, the best idea is focusing on the convergence of the system output and meanwhile guaranteeing boundedness of other states (internal signals). Consequently, the control law dimension and its implementation costs are reduced. The proposed single-loop controller design strategy is founded on the actuators' electrical subsystem. The closed-loop controlled system is established as BIBO stable, and the link position-tracking errors are uniformly bounded. Experimental implementations support the viability of the suggested theoretical results.