Abstract
This study describes online mass estimation and adaptive control of a multi-rotor helicopter. The altitude model is calculated in order to learn the relationship between the altitude and the mass. Here, motor dynamics and air resistance, and the approximation of the rotor frequency and the thrust power relationship are considered. The estimation approach is a nonlinear filtering problem for a augmented state space. The mass of a helicopter is augmented to the state vector, and the augmented state vector forms a nonlinear system. As a filter for a nonlinear system, EKF (Extended kalman Filter) and UKF (Unscented Kalman Filter) are compared. A simulation and offline data processing are done to verify the efficiency of the estimations. Furthermore, Self-Tuning PID Controller is designed. Once PID gains are tuned, Self-Tuning PID Controller tunes these gains automatically in response to the mass variation. This tuning algorithm is based on a characteristic equation of the feedback system. Altitude control simulation is done to verify the controller performance. In this simulation, mass of a UAV varies sharply. The controlled altitude and automatically tuned gains are shown.
