Design, modeling and control of quadrotor type unmanned aerial vehicle

This thesis considers design, modeling and control of quadrotor type unmannedaerial vehicle. The derivation of the dynamic model of the vehicle is performed byaid of Newton-Euler formulation and coefficients of the model are calculated basedon the real time model. The dynamic behavior of the actuator propulsion underlimited power conditions, which is a very important issue in unmanned aerialvehicles (UAVs), is particularly investigated. The aim of the propulsion modeling isto eliminate the modulation and decrease in battery voltage due to consumption ofpower by brushless direct current (BLDC) motor and incorporate the characteristicsof propulsion into quadrotor dynamics. Transient and steady state behavior of thepropulsion is represented by artificial neural network based approach whoseparameters are adjusted by Levenberg-Marquardt optimization technique. Thecontrol applications on quadrotor are achieved by two different forms, which aresimulation and real time studies. The accuracy of control approaches with derivedquadrotor dynamic model and neural network based propulsion model are justifiedvia simulation studies before real time applications. Proportional integral andderivative (PID) control scheme, sliding mode control (SMC), backsteppingtechnique and feedback linearization control are discussed, respectively, Acomparison of the approaches is presented in terms of the tracking precision,applicability of control signals and the qualities of the transient response. In the realtime studies, PID control and SMC techniques discussed in simulations are chosen todrive the vehicle dynamics to desired trajectories while sustaining stability. Thesolutions to problems which are encountered in real time studies are also discussed.