Calculate Goal Configuration for the Mobile Manipulator with Redundant DOF Based on Extended Manipulability Inversion

This paper deals with the problem of calculating joint space goal configuration for the mobile manipulator (MM) with redundant degrees of freedom(DOF) when performing grasping tasks. In real-world applications, the end effector (EE) goal state of MM is typically specified in the cartesian space while the motion planner is operating in the joint space. Therefore, it is necessary to derive the joint space goal configuration from the cartesian space goal state of EE. Due to the usual kinematic redundancy of MM, transforming this EE goal state to a goal configuration using inverse kinematics will result in infinite solutions. In this paper, the kinematic model of MM is constructed firstly, then the extended manipulability of MM is defined considering factors including manipulability, joint limit and self-collision avoidance. The Inverse Extended Manipulability Distribution which represents the extended manipulability of all possible base poses in the EE coordinate system is then built offline. Finally, by selecting the base pose with the highest extended manipulability through an online query process and solving the inverse kinematics of the arm, the goal configuration of MM is fully obtained. The effectiveness of the proposed algorithm is verified by simulation experiments in the end.