A twist decomposition approach to solve functionally-redundant serial manipulators.

摘要

This thesis proposes a new redundant-resolution approach to solve robotic tasks requiring less than six degrees-of-freedom (DOF). This approach is called Twist Decomposition Approach. Instead of projecting an optimization criterion onto the null space of the Jacobian matrix as most of the redundancy-resolution schemes do, twist decomposition approach firstly decomposes the Cartesian twist of the end-effector into two suitable subspaces; one being the subspace where the main task undergoes, while the other one being the redundant subspace. The redundancy resolution is optimized on the redundant subspace of the twist. Our approach could be applied to all redundant tasks requiring less than six DOF, regardless of how many DOF the manipulator has. (Abstract shortened by UMI.)