Gestural transmission of tasking information to an airborne UAV

Abstract

A system is presented that enables an authorized person on ground to transmit mission information to an airborne UAV within line of sight by using gestural expressions of both arms without the need for additional devices on ground. A miniaturized processing board with a discrete GPU is used to detect the body movements via a high resolution onboard camera and to translate them into relevant tasking information. Individual task elements are transmitted consecutively, including numerical and non-numerical information. A context aware gesture recognition approach is implemented to enable the reuse of gestures for different contexts in order to maintain a small gesture set. The system further features a bidirectional communication which allows to dispatch visual feedbacks and to query missing information visually via a LED matrix. Two experiments with different briefing contents in a static and dynamic setup have been conducted to proof the feasibility under real-life conditions.