Design Waypoint Navigation for Quadrotor UAV with simulation

OData support
Dr. Stumpf Péter Pál
Department of Automation and Applied Informatics

The topic of my project is very appealing field among researchers, engineer students and hobbyist: The UAVs (Unmanned Arial Vehicle) or with a more popular expression: ‘drones’. The most popular drone of nowadays, the Quadrotor UAV or simply quadrocopter, can be easily purchased for low price. The controlling of such a system is a nice example of modern controlling technics (fast sampling time, nonlinear control theory, state estimation). Describing the dynamical behavior of a quadrocopter is itself a nice mechanical problem, and controlling its rotors using a gyro sensor and an accelerometer is such a task where I could use almost all the knowledge which I have learned during my bachelor mechatronic engineer studies.

During this project an attitude controller (so called flight controller), and waypoint navigation were developed by Software in the Loop Simulation. This means that firstly the mathematical model of the aircraft was derived and implemented in Matlab/Simulink environment to investigate the nonlinear behavior of the system. Then, by using the simulation model, a flight controller algorithm was designed and tuned to control the attitude of the quadrocopter. A typical commercial remotely controlled quadrocopter contains similar controller. The feedback is the signal of the gyro-sensor. During the simulation some random noise added to the sensor signals to make the simulation more realistic.

Furthermore a waypoint navigation method was also developed, which –together with a GPS- can be functioned as an autopilot. The waypoint navigation means that the vehicle follows a path declared by points in the 3 dimensional space. Between the waypoints the aircraft tries to follow the shortest path, which is the linear. This algorithm contains 3 PID controllers for each axis, where the feedback signal is obtained from the accelerometer.

Using this simulation the constants of controllers can be tuned for any quadrocopter if the technical data is given or can be measured, so -as a further work- the feasibility of the designed controllers is possible.


Please sign in to download the files of this thesis.