This Bachelor's Thesis studies the implementation of machine learning for the preliminary design of the autopilot of a MAV apping wing vehicle. For this purpose, reinforcement learning is applied in this study through the development of a Q-learning algorithm. Due to the complexity that a apping wing vehicle presents, the aim of this project is to design an autopilot that is able to reach longitudinal control of a vehicle whose aerodynamic model is almost unknown. In order to develop a tool that accounts for complete longitudinal control, intermediate autopilot designs have been carried out so that the introduction of degrees of freedom to the problem has been gradual. Throughout the project, there is special concern about the development of a time and cost effective tool. ; Ingeniería Aeroespacial (Plan 2010)
Preliminary design of the autopilot of an autonomous apping wing micro air vehicle
2018-07-01
Hochschulschrift
Elektronische Ressource
Englisch
DDC: | 629 |
Homing and Docking Autopilot Design for Autonomous Underwater Vehicle
Trans Tech Publications | 2014
|Study of Low Cost Micro Autopilot for Fixed-Wing UAV
Trans Tech Publications | 2011
|Heading Hold Autopilot Design for Fixed Wing Aircraft
AIAA | 2018
|