抄録
This study describes online system identification of an altitude model and adaptive control of a multi-rotor helicopter. The identification approach is a nonlinear filtering problem for a augmented state space. As a system identification algorithm, EKF (Extended Kalman Filter) and UKF (Unscented Kalman Filter) are compared. Furthermore, Self-Tuning PID Controller is designed. Once PID gains are tuned, Self-Tuning PID Controller tunes these gains automatically in response to the system variation. This tuning algorithm is based on a characteristic equation of the feedback system. A simulation and offline data processing are done to verify the efficiency of the algorithm.
