ReDroId: Remote Drone Identification Based on Visual RSA SecurID Tokens

In an “open skies” era in which drones are flying among us, a new question arises: how can we distinguish between a friend drone (e.g., carrying a pizza delivery) and a foe drone (e.g., carrying a bomb) in areas that allow drone flights if both drones are identical? In this paper, we propose an authentication scheme that can help an authenticatee distinguish between a foe and friend drone. The authentication is implemented by optically modulating an RSA token using flashing lights which are mounted on the drone. Using a video camera, a trained machine learning model recognizes the flashing lights from the drone and converts them to a binary code which is used to authenticate the drone's identity.