Modeling of a quadrotor unmanned aerial vehicle, lyapunov based and backstepping controller design under gaussian noise and parameter uncertainty

Nowadays, unmanned aerial vehicles are used in a very wide area. Combating terrorism, reconnaissance and surveillance, fighting fire and natural disasters, search and rescue, mapping, mineral exploration, scientific research, cargo transport, organ transport, agricultural spraying, photography, hobby and sportive flights are among the areas of use of unmanned aerial vehicles. Unmanned aerial vehicles are more advantageous compared to manned aircraft due to their lower production cost, not carrying pilots, and being able to perform harder maneuvers thanks to their small size. Unmanned aerial vehicles can be considered in two categories: winged unmanned aerial vehicles and vertical take-off and landing unmanned aerial vehicles. Unmanned aerial vehicles that can take-off and land vertically do not need a runway during take-off and landing, they have the ability to hang in the air, rotate around their own axis, flip in the air and perform harsh maneuvers. These features make unmanned aerial vehicles capable of vertical take-off and landing more advantageous compared to fixed-wing unmanned aerial vehicles. Quadrotor unmanned aerial vehicles are four propeller vehicles that can take off and land vertically. The system dynamics of the quadrotor unmanned aerial vehicle are similar to the motion of a rigid body in space with 6 degrees of freedom. These 6 degrees of freedom consist of roll, pitch and yaw angles, known as Euler angles, which enable the formation of angular velocities, and x, y, z axes, which define linear movements in three-dimensional space. Within the scope of this thesis work, first of all, a nonlinear model of the quadrotor unmanned aerial vehicle was created. This created model was implemented in computer environment using MATLAB/Simulink software. Then, different controller designs were developed to control the altitude, roll, pitch and yaw angles of the quadrotor. Within the scope of the thesis work, an attempt was made to obtain a robust controller design that would work in the most durable way against all these difficult conditions. First of all, PID controller design was carried out due to its widespread use in the industry. For this controller, optimum controller coefficients were obtained by using the PID Tuning method. Then, an attempt was made to obtain a more robust controller approach than the PID controller commonly used in the industry. For this purpose, firstly, a Lyapunov based controller design based on the Lyapunov stability theorem was carried out. Then, a backstepping controller design, which is a recursive adaptation of the Lyapunov stability theorem, was carried out. The designed controllers were tested under normal conditions, white Gaussian noise, parameter uncertainty, simultaneous parameter uncertainty and white Gaussian noise, pink noise, brown noise, blue noise, purple noise and sinusoidal disturbance. In addition, triangle reference and sinusoidal reference tracking simulations were made to test the hard maneuvering capabilities of the controllers. The rise time, overshoot and settling time data of the designed controllers were compared. The data obtained showed that the backstepping controller had the least overshoot and the shortest settling time under all conditions. Thus, as a result of the simulations and the data obtained, it has been proven that the backstepping controller design is more robust than the classical PID controller and the Lyapunov based controller.