Quadrotor´s trajectory tracking control using monocular vision navigation

Precise trajectory tracking control for a quadcopter aerial vehicle using only a monocular camera and inertial sensing for localization is introduced in this work. Two different control laws based on linear behavior are proposed to follow desired trajectories. The algorithms are validated in real-time in the low-cost Parrot AR.Drone vehicle. Time-scale separation of the traslational and rotational dynamics allows to design controllers for the position of the quadrotor by giving desired reference roll and pitch angles to the internal autopilot. Localization is achieved via monocular SLAM (Simultaneous Localization and Mapping) and a suitable state estimator, both available as open-source code. The experimental results depicted in some graphs show the good behavior of the closed-loop system.