Longitudinal System Identification of an Avian-Type UAV Considering Characteristics of Actuator

To develop autonomous flapping aircraft, basic mathematical modeling is essential to solve the problem of “uncertainty models” in biomimetic flight dynamics. This study identified the longitudinal linear model of an ornithopter via automated flight tests and sensors onboard the ornithopter to measure angular rates, Euler angles, and total velocity. For accurate flight tests, automated signal input was designed for elevator deflection: doublet and multistep 3211 maneuver. Furthermore, because the flapping motion of the ornithopter’s main wings generates oscillations during cruise flight, fast Fourier transform is used to analyze flight data in the frequency domain, and a Butterworth filter is used to filter out the flapping motion from the data. The characteristics of the actuator are then analyzed using a motion capture camera and applied to the system identification. The structure of the ornithopter linear model is found to be similar to that of a fixed-wing aircraft, which has a periodic oscillation. The results from the flight tests and analysis manifest that an ornithopter has unstable characteristics due to a single right-half plane real pole.