A study on indoor positioning of an open-source drone using AR markers

When drones fly autonomously, GPS is basically used to establish navigation and safety functions. However, the above functions cannot be used in non-GPS environments such as indoor flight. This study focuses on indoor drone positioning using AR markers and aims to clarify its accuracy with a self-localization algorithm based on open-source platform. In the experimental evaluation, the drone flies autonomously by detecting AR markers with a drone-mounted camera, and its position coordinate data are collected to evaluate the accuracy of the self-localization. In particular, this paper evaluates the contribution of the position estimation by the AR markers to the self-localization computed by the flight controller, and shows that the self-localization accuracy is highly dependent on the position estimation by the AR markers, and its accuracy is about 1 to 2 cm.