Autonomous Navigation in Partially Observable Environments Using Hierarchical Q-Learning
Ye Zhou (TU Delft - Control & Simulation)
EJ Kampen (TU Delft - Control & Simulation)
Q. P. Chu (TU Delft - Control & Simulation)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
Flapping-wing MAVs represent an attractive alternative to conventional designs with rotary wings, since they promise a much higher efficiency in forward flight. However, further insight into the flapping-wing aerodynamics is still needed to get closer to the flight performance observed in natural fliers. Here we present the first step necessary to perform a flow visualization study of the air around the flapping wings of a DelFly II MAV in-flight: a precision position control of flight in a wind-tunnel. We propose a hierarchical control scheme implemented in the open-source Paparazzi UAV autopilot software. Using a decoupling, combined feed-forward and feed-back control approach as a core, we were able to achieve a precision of 2:5 cm for several seconds, which is much beyond the requirements for a time resolved stereo PIV technique.