Optimal kinematic control of humanoid arms with offset wrist

Abstract

This paper proposes the optimal inverse kinematic solutions of a 7R humanoid arm using the reconfiguration method and the self-optimizing method. Firstly, based on analyses of the humanoid arm with offset wrist, the reconfiguration method is proposed. Given the specified hand pose in Cartesian space, a closed-loop robot is developed by virtually connecting two ends of the humanoid arm. And then, the equivalent robot arm with a spherical wrist can be obtained by cutting open the kinematic chain at another link. And closed-form solutions of the humanoid arm are determined based on the seventh joint parameterization. Secondly, in order to realize kinematic optimizations of joint limits avoidance, the self-optimizing method is proposed based on the virtual force. The joint movable range is divided into the dangerous zone and the safe zone, and the evaluation of solutions in the dangerous zone is equivalent to the virtual force. Driven by the virtual force, the kinematic optimizations are performed using the iterative algorithm. Finally, kinematic simulations demonstrate that the virtual reconfiguration method is valid for inverse kinematics of humanoid arms with offset wrist, and kinematic optimizations are performed efficiently using the self-optimizing method.