End-effector pose estimation of the Monash Epicyclic-Parallel Manipulator through the visual observation of its legs

Abstract

Past research works have shown that it was possible to evaluate the end-effector pose of parallel robots by vision. First of all, it was proposed to directly observed the end-effector. However, this observation may be not possible (e.g. in the case of a haptic device for which the end-effector may be hidden by the user hand). Therefore, it has been proposed another type of end-effector pose estimation based on the observation of the directions of the legs. Even interesting results were obtained, this method is not suitable for some particular parallel robot families (e.g. the Monash Epicyclic-Parallel Manipulator, MEPaM). This paper proposes a new approach for the estimation of the end-effector pose: by observing the mechanism legs, it is possible to extract the Plücker coordinates of their lines and determine the end-effector pose. The new end-effector pose estimation method is applied to the MEPaM. All results are validated on a MEPaM simulator created using ADAMS/Controls and interfaced with Matlab/Simulink.