Intermittent Remote Visual Servoing in a Heterogeneous Robotic Team

Abstract

Unmanned aerial vehicles (UAVs) are becoming common technology used for various purposes. Most of the applications constitute monitoring and surveillance. In addition to being passive actors in search and rescue missions, UAVs take on more active roles in the form of aerial manipulators capable of physical interaction with the environment. In this paper, we present the collaborative approach between UAV and robot manipulator, where the robot manipulator plays the role of the visual observer. we also present a UAV control based on visual servoing, a state machine for suction-based object grasping in a GNSS-denied scenario, and study the effect of different message drop probabilities on the system performance. The presented scenario is one of the tasks of the MBZIRC competition. Results are presented in the official contest simulator, Ignition Gazebo.