Modeliranje in krmiljenje kvadrokopterja z uporabo LQR metode v programskem okolju Simulink

Bratuša, Urška (avtor), Podržaj, Primož (mentor)

V okviru magistrske naloge smo modelirali in krmilili kvadrokopter z uporabo LQR metode v programskem okolju Simulink. Model kvadrokopterja smo linearizirali okrog točke lebdenja in ga predstavili v prostoru stanj, nato pa s pomočjo LQR metode določili matriko ojačanj K za linearni model. Enak pristop smo uporabili tudi za nelinearni model. Oba modela smo analizirali z vidika odziva na skočno funkcijo. Na nelinearnem modelu smo primerjali odziv sistema z uporabo LQR metode in PID krmilnika. Krmiljenje je vključevalo rotacije okrog osi x, y in z ter translacijo v z – smeri. Poleg tega smo izvedli identifikacijo ključnih parametrov za modeliranje kvadrokopterja s pomočjo sistema za zajemanje podatkov. Ugotovili smo, da LQR metoda, zasnovana za linearne modele, učinkovito deluje tudi na nelinearnih modelih in da zagotavlja odziv brez prenihaja v primerjavi s PID krmilnikom

• Jezik: Slovenski jezik
• Ključne besede: kvadrokopter, LQR algoritem, PID krmilnik, nelinearen model, Simulink
• Vrsta gradiva: Magistrsko delo/naloga
• Organizacija: FS - Fakulteta za strojništvo
• Leto izida: 2024
• PID: 20.500.12556/RUL-161648
• Datum objave v RUL: 13.09.2024

Sekundarni jezik

• Jezik: Angleški jezik
• Naslov: Modeling and control of a quadcopter using the LQR method in Simulink

Izvleček

In the master's thesis, we modeled and controlled a quadcopter using the LQR method in the Simulink software environment. The quadcopter model was linearized around the hovering point and represented in state space, then the gain matrix K for the linear model was determined using the LQR method. The same approach was applied to the nonlinear model. Both models were analyzed in terms of step response. In the nonlinear model, we compared the system response using the LQR method and the PID controller. Control included rotations around the x, y, and z axes and translation in the z – direction. Additionally, we identified key parameters for modeling the quadcopter using a data acquisition system. We found that the LQR method, designed for linear models, also works on nonlinear models and provides a response without overshoot compared to the PID controller.

• Ključne besede: quadcopter, LQR algorithm, PID controller, nonlinear model, Simulink