Experimental Safety Analysis of R-Min, an Underactuated Parallel Robot

Abstract

The R-Min robot is an intrinsically safe parallel manipulator dedicated to pick-and-place operations. The proposed architecture is based on a five-bar mechanism, with additional passive joints in order to obtain a planar sevenbar mechanism with two degrees of underactuation, allowing the robot to reconfigure in case of a collision. A preload bar is added between the base and the endeffector to constrain the additional degrees of freedom. This article presents an analysis of the workspace and of the safety performances of the R-Min robot, and it compares them with those of the five-bar mechanism, in order to evaluate the benefits of introducing underactuation in a parallel architecture to obtain intrinsically safer robots. The geometrico-static model of the R-Min robot is formulated as an optimisation problem. The direct and inverse kinemato-static models are derived from the geometricostatic model and they allow to express the singularity conditions of the R-Min robot. An analysis of the singularity loci is carried out among the robot's workspace. A controller based on the dynamic model is proposed and experimentally validated on a prototype of the R-Min robot. Finally, the safety performances of the R-Min robot are evaluated experimentally and they are compared with that of an equivalent five-bar mechanism, using the maximum impact force as a safety criteria in accordance with recent international standards.

Dynamics and Computed Torque Control Stability of an Under-Actuated Tendon-Driven Manipulator

“…We apply the computed torque control that linearizes the dynamic equation [9] to underactuated systems [10]. This control has been applied on other under-actuated manipulators like the parallel robot of [11] or the cable-driven parallel robot of [12]. As observed in [13] for under-actuated cable-driven parallel robots, it will be seen that the computed torque control can be unstable.…”