The Unmanned Aerial Vehicles (UAV) escape from missions which were constantly limited because of high risk of human operation, and now have greater maneuverability. The unlimited potential energy of UAV to fulfill themselves in newly required mission achievements come from its reduced need for human intervention, increased performance range and capabilities, extended operation life and de-creased costs. The accurate system configuration of the UAV in re-quested mission environments is demanded to design appropriate controllers for autonomous flight and the key point is to identify the parameters of the system.
The traditional approaches used in the case of aircraft are maxi-mum likelihood, linear regression and Extended Kalman Filter (EKF). In this thesis, the Unscented Kalman Filter (UKF) is proposed as a new parameter estimation method for helicopter model complementing flaws of previously introduced schemes. The UKF especially has higher accuracy when it is used in estimation of nonlinear model parameters and the comparison of its performances with EKF through numerical simulations is presented. The small scaled unmanned Rotary UAV (RUAV) system is constructed and based upon numerical simulations a proposed estimation method is applied.
The RUAV has flight control computer, sensors, power supply system, etc. The flight data needed to estimate parameters of the system is acquired through flight test. The UKF application to real flight data and its results are also presented.