Abstract
This paper presents a new manipulation theory for controlling compliant motions of a robotic manipulator. In previous closed loop control methods, both direct kinematics and inverse kinematics of a manipulator must be resolved to convert feedback force and position data from Cartesian space to joint space. However, in many cases, the solution of direct kinematics in a parallel manipulator or the solution of inverse kinematics in a serial manipulator is not easily available. In this study, the force and position data are packed into one set of “motion feedback,” by replacing the force errors with virtual motion quantities, or one set of “force feedback,” by replacing motion errors with virtual force quantities. The joint torques are adjusted based on this combined feedback package. Since only Jacobian of direct kinematics or Jacobian of inverse kinematics is used, the computational complexity is reduced significantly, and the control scheme is more stable at or near singular manipulator configurations. Furthermore, the complexities and oddities associated with hybrid control, such as nonuniformity of the space matrix and incompatibility of simultaneous position and force control in the same direction are circumvented. The applications of this theory are demonstrated in simulation experiments with both serial and parallel manipulators.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call