Landing Simulated Unmanned Aerial Vehicles With Proxy Manipulation Gestures
Marvin Brenner, Peter Stütz · 2024
Gesture-based interfaces for human-drone interaction enable novice users without technical background to access higher-level functions of highly automated unmanned aerial vehicles. Directly controlling the flight with one's own hands represents an intuitive control method but is difficult to realize in the field without additional equipment. Image-based tracking of the operator's hands is becoming increasingly accurate and offers a possible low-cost solution through gesture recognition. Because gesture vocabularies are use case dependent, it is difficult to find a standardised interaction method. By providing a modular, experimental framework for testing gesture-based interfaces in various simulation environments, this work aims to reduce the amount of time from conceptualizing the gesture vocabulary to deploying it in a real flight test. It is demonstrated by comparing two hand tracking strategies for a 3D proxy manipulation gesture, how novel gesture interfaces and recognition methods can be quickly deployed in simulation.