Extendable Flight System for Commercial UAVs on ROS

In this work, we propose a flexible and extendable framework which is integrated into the Robot Operating System (ROS) on an onboard companion computer for DJI-autopilots based Unmanned Aerial Vehicles (UAVs). Specifically, considering that the DJI-autopilots do not directly support external position data except those obtained from GPS or DJI Guidance, to fully take advantages of the stability and reliability of DJI-autopilots, a high-level state estimator is designed to provide the position and velocity of the UAV. And an outer-loop geometric tracking controller that can compute the desired thrust and attitude is introduced to accomplish the flight tasks generated by a guidance module. Furthermore, this framework is convenient to be implemented because of well-defined communication interfaces on ROS. Finally, both Hardware-In-The-Loop (HIL) simulations and actual experiments are presented to demonstrate the effectiveness of the proposed framework.