Control of the flight of a small quadrotor using gestural interface

Abstract

A new approach using gestures and visual computing techniques to control unmanned aerial vehicles is presented. The goal is to control automatically and in semi-autonomous mode the flight of a quad-rotor rotorcraft in indoor environments characterized with the absence of possibility to receive GPS data. Human operator can control the implementation of various maneuvers during the flight of the rotorcraft via specific gestures and body postures. In this way a more-effective interface with the controlled aerial robot is created. For its practical implementation a Parrot AR. Drone quadrotor and a Microsoft Kinect sensor have been used.