Humanoid reaching posture selection under geometric constraint considering manipulabiltiy

Abstract

The goal in this research was to develop an optimization algorithm for a humanoid to enable it to reach a point. Simply stretching its arms, which in a humanoid are shorter than its body and legs, is not sufficient to reach an object located at some distance away or on the ground. Therefore, reachability has to be ensured by a combination of motions including kneeling and orienting the pelvis. However, many posture selection options exist because of the redundancy of a humanoid. In this research, we focused on the optimization of the posture of a humanoid that is reaching toward a point. The global optimal solution of the problem is easily obtained by searching method. But it needs too much heavy computing time. For that reason, it is evident that the optimization problem should be simplified. First, the proposed posture selection simplify the pelvis-related tasks as a kinematic model which of joints are height, pitching and yawing. And the objective function of the optimization problem is changed from full manipulability to forearm manipulability. Since the orientation singularity is escapable, it is evident that the forearm manipulability is considered for the redundant DRC-HUBO+ arm. Furthermore, the forearm manipulability can be projected to the distance from shoulder position to wrist position and the optimal distance is obtained. The optimal distance is defined as a geometric constraint for the posture optimization problem. The optimization problem minimizes the displacement from the initial posture to the reaching posture while the optimal distance is satisfied. The geometric constraint can be used to maximize or minimize other performance such as force, energy etc. The feasibility is demonstrated by many experiments of the motion planning for removal debris blocking the path that require the reaching posture for random reaching point.