Biomimetic joint/task space hybrid adaptive control for bimanual robotic manipulation

Abstract

In this paper, we study the adaptive control of a bimanual manipulator moving a dynamic object through a trajectory. Impedance and force are adapted online using a novel biomimetic algorithm that minimises both tracking error and control effort, as observed in humans. On top of our previous work of impedance and force adaptation, a new task space/joint-space hybrid control scheme is developed. The task space controller adapts end-point impedance, to compensate for interactive dynamics, and the joint-space controller adapts the impedance to improve robustness against external disturbances. Extensive simulations demonstrate the efficiency of the developed adaptive motion controller. The results show that the proposed hybrid controller can perform well under large disturbance conditions while minimising control effort and tracking error.