Velocity-free image-based control of Unmanned Aerial Vehicles

Controlling aerial vehicles motion considerably relies on the measure of their velocity. Yet, obtaining the translational part of such information from solely the on-board sensors is still an open issue to envisage fully autonomous applications. In this paper, we present a nonlinear observer based on the images from a single on-board camera to on-line estimate the translational velocity. Spherical image coordinates are adopted. Then, through Backstepping design, the result is exploited in a visual servo controller, thus endowing the vehicle with the capability of fully autonomously positioning on visual targets. The stability and convergence of the closed-loop system are established through Lyapunov synthesis. By means of computer simulations, the validity and robustness of the proposed system are shown.