Abstract
This study elaborates on system identification of flight characteristics of small unmanned aerial vehicles (UAVs), which are defined as fixed-wing aircraft with wingspan of approximately 1 m and gross weight of 1 kg. Although system identification is well-established as a method to acquire flight characteristics of large aircraft, few cases apply this method to small UAVs. Therefore, the goal is to examine whether techniques for large aircraft are applicable to small UAVs. This study focuses on two obstacles: 1) An avionics device that obtains the required data to perform system identification must be far lighter and smaller than the existing ones for installation into small UAVs, and 2) an appropriate analytical method must be used to obtain accurate results because small UAV flight differs significantly from large aircraft flight, such as in terms of low gust resistance. For the first problem, new small and lightweight avionics utilizing micro electro-mechanical system (MEMS) sensors are proposed. The second obstacle is studied by applying four sophisticated analytical methods and the results show that unscented Kalman filter (UKF) is superior to recursive least square (RLS), Fourier transform regression (FTR), and filter error method (FEM).
