Hybrid Inverse Kinematics for a 7-DOF Manipulator Handling Joint Limits and Workspace Constraints

A manipulator with a human-like range of motion combined with adequate control strategies can increase the autonomy of humanoid robots. If we look at the motion patterns of human beings during manipulation, we see that they often combine trunk and manipulator motion to grasp objects lying outside the workspace of the manipulator itself. In this work, we introduce an approach for a hybrid inverse kinematics algorithm with a focus on avoiding joint and workspace limitations. We combine a fully analytical inverse kinematics algorithm for a redundant 7-DOF manipulator with the inverse kinematics of the robot’s trunk. To handle the workspace limitations, if a desired pose lies outside the workspace, we combine the robot’s trunk motion with the manipulator motion. We demonstrate the performance of our approach with the humanoid robot Sweaty, playing fully autonomously a chess game against a human while picking up chess pieces lying outside the workspace of its manipulator by a combined trunk and manipulator motion.