Model derivation, attitude control and Kalman filter estimation of a quadcopter

Recently unmanned aerial vehicles have become one of the most interesting research topics among scientists. Although researchers are very concerned in this area, they generally use the same dynamical plant equations. Simulations using that ready-to-use plant equations are common methods in order to apply control or optimization theories. But, beyond simulations, in real time control of these systems may yield some unknown results that drives the plant from optimum state and/or stability. Starting from this point of view, we first prepare basic test bench for our quadcopter. Then, take some observation data includes motor parameters and predefined attitude angles. By help of measurement phase, we derived differential equations for our quadcopter. Finally, a continuous time steady-space model is obtained from these equations. In Matlab, a linearization procedure is done under some constraints such as maximum allowable angle rates along hover position. Then, a PD type classical controller is designed for both altitude and attitude control of the quadcopter. A Gaussian distributed noise is added over the controlled states of the quadcopter as disturbance of physical phenomena. Finally, a Kalman filter is designed and implemented in order to estimate the states of the noisy signals to achieve a reliable control of the quadcopter.