Modeling and control of a quadrotor

This thesis deals with the modeling and control of an Unmanned Aerial Vehicle called as 'quadrotor'. The quadrotor which is capable of vertically taking off and landing, does its basic movements by adjusting the rotational speeds of propellers fixed on four rotors which are mounted on a crossed or plus structured arms. These basic movements are hovering, rising, pitching, rolling and yawing. The quadrotor can go forward, backward, to the left or to the right by generating moments at its center of gravity. These moments occur when the propeller pairs rotate at different speeds. In this work, the quadrotor model is derived in order to describe these movements mathematically. The six degree of freedom model in body frame is provided by using the Newton-Euler equations The linear control methods such as PID and Linear Quadratic Regulator and the nonlinear control methods such as backstepping and sliding mode control are preferred to control the quadrotor which has a nonlinear model. All the controller designs are agreed with the pre-determined constraints. The performance of the control methods are observed by analyzing the step responses and the trajectory following success given with different scenarios. A robust system is insensitive to uncertainties. In this work the robustness of the designed control methods are tested by adding disturbance and noisy measurements to the system. The simulation results show that the quadrotor can be controlled by any methods given in this thesis except the PI controller. The most successful control method in following the trajectories is the Linear Quadratic Regulator designed over a single model it is also the most robust method. But it exhibits more control efforts than others. PD control is found to be more difficult to restabilize the quadrotor when its exposure to a disturbance compared to the other methods. With this study, it is aimed at determining which controller is more suitable to control a quadrotor.