Trajectory tracking through overwhelming control of human vertebrae as a hybrid manipulator

Abstract

Industrial robots, generally, are composed of architecture with serial manipulators having large workspace but these suffer from several drawbacks, such as low accuracy and low load-carrying capacity. Similarly, the parallel architecture robots have better accuracy but small workspace. Therefore, the combination of serial and parallel chain architectures has become attractive as hybrid manipulators over the last two decades. A planar hybrid manipulator is proposed here and this is made by placing a planar parallel manipulator, having three degrees of freedom, each over the other in series so as to increase its load-carrying capacity and then it is modeled with the help of the bond graph approach. The human vertebrae are considered as a hybrid manipulator in this paper. The translation of the human vertebrae model on the frontal plane and lateral bending in the right- and left-hand directions are proposed here to increase the flexibility of the body part of humanoid robots used in industries and sports so that they can execute spine bending activities. The forward and inverse models for the manipulator are presented and the system inversion is carried out through the overwhelming controller. Finally, the simulation results for trajectory tracking of left-hand lateral bending in the frontal plane is done with or without the consideration of leg inertia and conclusions are presented to show that response follows the command within acceptable limits.