We present an enhanced autonomous indoor navigation system for a stock quadcopter drone where all navigation commands are derived off-board on a base station. The base station processes the video stream transmitted from a forward-facing camera on the drone to determine the drone's physical disposition and trajectory in building hallways to derive steering commands that are sent to the drone. Off-board processing and the lack of on-board sensors for localizing the drone permits standard mid-range quadcopters to be used and conserves the limited power source on the quadcopter. We introduce improved and new techniques, compared to our prototype system [1], to maintain stable flights, estimate distance to hallway intersections and describe algorithms to stop the drone ahead of time and turn correctly at intersections.
Autonomous indoor navigation of a stock quadcopter with off-board control
01.10.2017
389219 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Autonomous Navigation of a Quadcopter in Indoor Environments
TIBKAT | 2019
|Navigation and Guidance for Autonomous Quadcopter Drones Using Deep Learning on Indoor Corridors
BASE | 2022
|