Motion tracking and object manipulation of a hyper-redundant MDAMS under joint limits based on task-priority & relative Jacobian

Abstract

With the increasing demand for personal assistant robots, a hyper-redundant non-holonomic mobile dual­arm manipulator system (MDAMS) which interacts with human beings in a domestic environment is introduced and a virtual prototype is established. In order to further construct the real robots, a task priority and relative Jacobian matrix-based semi-decentralized motion tracking and object manipulation strategy is proposed in this paper. To facilitate the modeling process, the forward kinematics are constructed using Modified Denavit Hartenberg (MDH) method with coupling investigation among the non-holonomic mobile base (MB), upper moving body and two arms. Besides, a task priority and relative Jacobian matrix-based inverse kinematic solving method is designed to deal with the hyper- redundancy problem induced by MB and redundant arms. The relative motion between two end-effectors (EE) is defined by introducing a compact relative Jacobian matrix. Furthermore, the unit quaternion is employed to characterize the EE's orientation and a modulated weighted least norm technique is introduced to dampen joint position in case the weight matrix is too large. Moreover, a semi-decentralized control strategy is presented with a kinematic control law for the MB. Object dynamics are constructed and object manipulation is realized using the relative motion tracking strategy. Finally, a mathematical MDAMS model is built and several trajectory tracking simulations are realized to validate the proposed strategy.