Adaptive Quadratic Programming Approach for Inverse Kinematics in Free-Flying Space Robots

Abstract

This paper presents a novel whole-body inverse kinematics (IK) control framework for free-flying space robots (FFSRs) to address the challenges of tracking and capturing non-cooperative targets in orbit. By formulating the IK problem as a quadratic programming (QP) optimization, the proposed method integrates multiple objectives, including precise end-effector velocity tracking, manipulability maximization, and base reaction minimization, within a unified framework. The algorithm introduces adaptive weighting mechanisms that dynamically balance precision, stability, and redundancy utilization based on real-time tracking errors and task demands. In simulations involving a 10-DOF mobile manipulator, the proposed method maintains end-effector positional errors below 0.03 m at tumbling speeds up to 1 rad/s, and under 0.05 m at linear target speeds up to 1 m/s. Hardware experiments with an 8-DOF mobile manipulator further verify real-time feasibility, showing final pose errors under 0.01 m in close-range capture. These results demonstrate the method’s effectiveness for fault satellite capture and on-orbit maintenance, providing a viable solution for servicing malfunctioning satellites and ensuring the sustainability of space assets.

Keywords: Inverse Kinematics, Free-Flying Robot, Reaction Torque Optimization, Redundant manipulators