Deep Neural Network-Inspired Approach For Human Gesture-Triggered Control Actions Applied To Unmanned Aircraft Systems

Deep Neural Networks have positively impacted the development of autonomous navigation strategies, allowing the integration of object detection capabilities to the vast array of already existing control methods. In this project, a Deep Neural Network is implemented to distinguish particular hand gestures, as well as user's arm movement, in order to trigger a specific activity of an Unmanned Aircraft System. We present real-time experimental results of how this technique is applied, making use of a Natural User Interface to control a the autonomous aircraft.