A Practical Visual Servo Control for Aerial Manipulation Using a Spherical Projection Model

Abstract

This article addresses the problem of autonomous servoing control of an unmanned aerial manipulator with the capability of grasping target objects using computer vision. Specifically, a practical visual servo control using a spherical projection model is proposed. The aerial manipulator is an unmanned aerial vehicle equipped with a robotic arm that greatly increases the freedom and operational flexibility of the end-effector. However, it also increases the complexity of the kinematics, dynamics, and control design of the complete system. A novel passivity-like error equation of the image features is established by using the spherical camera geometry dynamics with an eye-in-hand configuration. To further improve the grasping performance, a task-priority control scheme is utilized with one main task and several subtasks, i.e., controlling the gripper position and orientation, vertically aligning the center of gravity, and avoiding the joint limitation. Simulation results are provided to illustrate and assess the performance of the proposed visual servo control. The practicability and effectiveness of autonomous aerial manipulation are well supported by the experimental results acquired through outdoor environments.