Date

2021-9-01

Author

Baskın, Mehmet

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

703
views

345
downloads

Cite This

In this thesis, design of a flight control system for an uncommon quadrotor aerial vehicle is discussed. This aerial vehicle consists of two counter-rotating big rotors on longitudinal axis to increase the lift capacity and flight endurance, and two counter-rotating small tilt rotors on lateral axis to stabilize the attitude. Firstly, full nonlinear dynamic model of this vehicle is obtained by using Newton-Euler formulation. Later, derived approximate linear model around hover is statically decoupled to simplify the flight dynamics. Resulting diagonal model is mainly based on physical principles and it does not include dynamics such as uncertain parameters, sensor delays, flexible modes of the structure and inexact decoupling. Therefore, a system identification method is used, and more accurate model is estimated from flight tests when this vehicle is stabilized by PI controllers with sufficient performance and stability margins. Secondly, relative importance of the parameters in the linear hover model is investigated. It is assumed that each parameter of the dynamic model has some uncertainty. Structured singular value sensitivity analysis determines the important parameters in terms of robust stability. Maximum tolerable uncertainties of parameters are also computed. High performance control design requires accurate model that is suitable for subsequent robust control design. Therefore, robust control criterion is considered during system identification step. System model is defined in terms of coprime factors which are identified by solving corresponding least squares problem using data-dependent orthonormal polynomials to achieve optimal numerical conditioning. Next, validation-based uncertainty modeling is used to compute the frequency dependent uncertainty upper bound. Using the uncertainty upper bound and specific coprime factorization of the stabilizing controller, robust-control-relevant model set is obtained which facilitates high performance robust controller synthesis. Finally, performances of the PI controller with sufficient stability margins and the designed robust controller are analyzed. Attitude stabilization and torque disturbance rejection performances are compared. Results of several flight experiments show attitude control performance improvement with the designed robust controller compared to the standard PI controller.

Subject Keywords

system identification, robust control, model uncertainty, attitude control, flight control, unmanned aerial vehicle, H-infinity, model validation

URI

https://hdl.handle.net/11511/93261

Collections

Graduate School of Natural and Applied Sciences, Thesis