Publication: Non-Linear Control for Autonomous Landing on Moving Platforms
Loading...
Identifiers
Publication date
2019-10-09
Defense date
2019-10-09
Authors
Advisors
Tutors
Journal Title
Journal ISSN
Volume Title
Publisher
Impact
Export
Abstract
En este proyecto, realizado como Trabajo de Fin de Grado para la Universidad Carlos
III de Madrid en el Laboratorio LSI, se pretende desarrollar un sistema de control
adecuado para conseguir aterrizar un UAV de tipo cuadricóptero en plataformas de aterrizje
móviles, con la intención de poder ser utilizado este sistema para diferentes tipos
de dron. Para ello se programará un nodo en C++ subscrito y publicando sobre topics de
MAVROS, que posteriormente enviará comandos de velocidad lineal a un control de más
bajo nivel del PX4 Autopilot.
Se controlará la posición del UAV utilizando un controlador PID de base que envíe
comandos de velocidad. Una vez ajustadas las ganancias de ese sistema de diferentes
modos y seleccionando las más adecuadas, se probarán distintos modos de mejorar y
optimizar el control.
Se implementará un control final que anexione el PID previamente diseñado, un seguimiento
de la plataforma móvil utilizando ecuaciones cinemáticas discretas basadas en
ecuaciones de Euler, un detector de suelo en el aterrizaje utilizando el controlador de bajo
nivel del Autopilot y una lógica por conjuntos difusos para conseguir ganancias para el
PID adaptables.
Los resultados y datos necesarios serán recogidos de la simulación en Gazebo del PX4
en SITL, utilizando como modelo un cuadricóptero “Iris”, probando el control en distintos
escenarios, como con plataforma de aterrizaje estática, con velocidad lineal o sinusoidal
y aplicando perturbaciones de fuerzas externas puntuales y fuerzas constantes durante un
tiempo dado.
This project was carried out as Final Grade Project for the University Carlos III of Madrid. In it, it is expected to develop a control system suitable for landing a quadcopter UAV on moving platforms, with the purpose of this system of being able of being used on different types of drones. To achieve that, a node in C++ will be programmed subscribing and publishing on topics from MAVROS, wich will, subsequently, send linear velocity commands to a lower level control in PX4 Autopilot. The position of the UAV will be controled using a PID controler as basis which sends velocity commands. Once its gain are tuned and adjusted using different ways and selected the proper ones, diverse ways of optimizing and improving this controller will be tested. A final control system will be implemented in order to assemble the PID previously designed, a moving platform follower which uses discrete cinematic equations based on Euler equations, a ground detector when landing which uses the lower level controller from the Autopilot and a fuzzy logic to achieve adaptive PID gains. Data and results will be obtained from the simulation in Gazebo and PX4 on SITL, using the quadcopter “Iris” as model, testing the control in different scenarios as using a static landing platform, using one with linear velocity or using sinusoidal trajectories and applying external disturbances like punctual or constant temporary forces.
This project was carried out as Final Grade Project for the University Carlos III of Madrid. In it, it is expected to develop a control system suitable for landing a quadcopter UAV on moving platforms, with the purpose of this system of being able of being used on different types of drones. To achieve that, a node in C++ will be programmed subscribing and publishing on topics from MAVROS, wich will, subsequently, send linear velocity commands to a lower level control in PX4 Autopilot. The position of the UAV will be controled using a PID controler as basis which sends velocity commands. Once its gain are tuned and adjusted using different ways and selected the proper ones, diverse ways of optimizing and improving this controller will be tested. A final control system will be implemented in order to assemble the PID previously designed, a moving platform follower which uses discrete cinematic equations based on Euler equations, a ground detector when landing which uses the lower level controller from the Autopilot and a fuzzy logic to achieve adaptive PID gains. Data and results will be obtained from the simulation in Gazebo and PX4 on SITL, using the quadcopter “Iris” as model, testing the control in different scenarios as using a static landing platform, using one with linear velocity or using sinusoidal trajectories and applying external disturbances like punctual or constant temporary forces.
Description
Keywords
UAV, UAVs, Aterrizaje autónomo, Cuadricópteros, PID, Lógica borrosa, IMU