Modelling and simulation of a quadrotor and altitude and position control using PD controller

A quadrotor is a X-shaped aerial vehicle with 4 rotors, capable of vertical landing and take-off. Thanks to high maneuverability, low and high flying speeds and simple mechanical structure, quadrotor's usage areas are constantly increasing. Many studies have been presented in the simulation environments and experiments for quadrotors and new studies are being carried out. Many control methods have been developed to control quadrotors. In this study, kinematic and dynamic equations of the quadrotor were obtained based on Newton-Euler principles. Using these equations, 2 simulation models of the quadrotor were designed. In the first simulation model, three linear controllers were designed, which are PD (Proportional-Derivative) position controller, PD (Proportional-Derivative) attitude controller and PI (Proportional-Integral) motor controller. In the second model, 4 linear controllers were designed, which are PD (Proportional-Derivative) position controller, cascade P (Proportional) velocity controller, PD (Proportional-Derivative) attitude controller and PI (Proportional-Integral) motor controller. Reference and measured position signals, measured linear velocities, measured Euler angles and position error signals of these two simulation models were presented and discussed. According to the simulation results, the quadrotor guaranteed trajectory at velocities up to 5 m/s. In the first and second simulation models, the settling time was between 3 and 4 seconds. In the 1st simulation model, there was no steady state error after the settling period. In the 2nd simulation model, after the settling time, steady state error occurred up to 5% of the linear velocity. In the simulations with the linear reference signal, the position error in the 2nd simulation model was less than the 1st simulation model during the settlement period. The only advantage of the second simulation model is that it reduces the error in the settling period in simulations where a linear reference signal is applied. Therefore, the 1st simulation model performed well with up to 5% error. Furthermore, it did not cause steady state error after the settling period. Using the 2nd simulation model only during the settling period and after the settlement period, using the 1st simulation model presented better results.