Multirotor systems have an ever increasing range of applications nowadays. They are becoming widely used in search and rescue missions, military and law enforcement, and are also useful for shooting videos in otherwise unreachable territories. A recent and upcoming area is combining many of them into an interacting group, building up a bird-like-flock of autonomous quadrotors. The latter is amongst the several future research opportunities they provide.
The long-term aim of this project was to completely put together a quadrotor from the very beginning including all the mechanical, electronical and informatical parts. An own hardware has the advantage of being mechanically flexible, besides control parameters and the control algorithm implemented on the onboard computer is also open to further changes. This way building your own hardware provides a better platform for further studies.
In my project, at first a QUAV’s dynamic model was investigated. Based on that, several separate (linear and nonlinear) motion control techniques and path tracking algorithms were examined and simulated - with the hope of testing them on a completely finished hardware once. Next, some of today’s most used trajectory planning techniques are mentioned, including those that are available in areas where Global Positioning System is not.
At last - in order to get better acquainted with the real control and dynamics - the behaviour of a dual rotor pendulum was investigated. The traditional attitude control of quadrotors uses Euler angles (yaw, roll, pitch) for keeping track of position. The last two movements (roll or pitch) can be approximated with a mostly symmetric rigid body - rotating about a fixed-axis. This system was modelled and stabilized (both in Matlab/Simulink environment and on real hardware), this way providing a good base for the quadrotor’s roll/pitch stabilization later.