Visual-Servo Autonomous Robotic Manipulators for Capturing Non-Cooperative Target

Abstract

This paper presents a methodology of vision-based pose and motion estimation of non-cooperative targets as well as a control scheme for robotic manipulators to perform autonomous capture of non-cooperative targets. A combination of photogrammetry and extended Kalman filter is proposed for real time state estimation of the non-cooperative target. Once the vision-based estimation is obtained, a real state of the target regarding to the global frame is calculated based on the transformation matrices of coordinate frames. So as to make a capture, a desired state of the end effector is defined in accordance with the real state of the target aforementioned, and further a corresponding desired state of the robotic manipulator is derived by inverse kinematics. Then a close-loop control scheme is adopted to drive the robot to the desired state previously obtained. Experiments have been designed and implemented on a custom built six degrees of freedom robotic manipulator with an eye-in-hand configuration. The experimental results demonstrated the feasibility and effectiveness of the proposed methodology and control scheme.