Motion planning and control of interactive humanoid robotic arms

Abstract

Humanoid robots are widely discussed in recent years. The motion planning and control of humanoid robots can be discussed based on mobility of platforms and manipulations of arms. In this paper, we propose a robotic arm which manipulation is analog to the motion of human’s upper extremities. The proposed robotic arm is designed as a seven degree-of-freedom configuration. To increase the interactivity with humans, a six-axis force sensor is attached on the wrist of the robot to capture the force applied on the robotic arm. Subsequently, the robotic arm is moved following the force applied on the wrist. In addition to the compliance of human’s motion, the robotic arm is capable of dynamically planning spatial trajectories for various straight lines, circles, or predefined paths. Especially, due to the structure of this seven degree-of-freedom robotic arm, we cannot find a unique solution for the inverse kinematics. In this work, we present a behavior based inverse kinematics approach to solve this problem in terms of the fuzzy reasoning. Various behaviors for a given spatial position or path, such as writing, pickup, etc., may result different inverse kinematic solution, and may generate different elbow trajectories as well. Therefore, the proposed robotic arm not only has similar structure to humans, but also represents similar behavior to humans. More specially, the compliance function makes this robotic arm possible to interact with humans. Consequently, a robotic arm with tendon driven architecture is demonstrated to validate the proposed motion planning and control approaches based on an ARM based controller.