Control of quadrotor UAV using variable disturbance observer-based strategy

Disturbance observer-based control is widely employed in control systems. It demonstrates the ability to estimate intricate disturbances for subsequent measurement, thereby enabling their compensation in the output of existing controllers. Nevertheless, disturbance observer-based control shows limited robustness when utilizing a fixed nominal model. Simultaneously, the fixed time constant of the Q-filter prevents the securement of filtering performance across various flight speed areas. This study aims to enhance robustness in both open- and closed-loop disturbance observer-based control by carefully selecting the Q-filter time constant through a comprehensive transfer function analysis. The transfer function characterizing the quadrotor unmanned aerial vehicle is derived from highly reliable dynamic modeling obtained through wind-tunnel testing, with the nominal model computed by the flight speed. We propose variable disturbance observer-based control by using a nominal model according to flight speed and selecting the Q-filter of disturbance observer-based control by considering the characteristics of the dynamic model, disturbance, and sensor noise. Numerical simulations and flight tests are conducted to compare the controller’s performance, focusing specifically on attitude tracking during a round-trip flight with a slung load.