External hybrid position/force control for kinematically redundant manipulators

摘要

Very few investigations on compliant motion control have been done for achieving tasks in constrained environmetns with kinematically redundant manipulators. In this paper, the hybrid position/force controller with an external structure is extended to redundant arms. This control scheme is combined with an iterative algorithm based on neural network for redundancy resolution. The redundant DOFs are not only used to meet addition constraints such as those classical equality constraints that parameterize the self-motion of the manipulator. They also make possible to meet inequality constraints that are imposed by the external environment and inevitably arise in constrained workspace. The objective function is then augmented with penalty functions.