Subspace-Based System Identification For A Hovering Tiltrotor Uav Attitude Dynamics

In recent years, the tiltrotor UAV (Unmanned Aerial Vehicle) attracts increasing concern. As progressively requirements are being imposed on its control systems, accurate dynamic modeling is becoming increasingly important. Subspace identification (SID) as a kind of the black-box identification method is suitable for multi-input -multi-output (MIMO) system and has been used in UAV system identification. But most of the previous literatures using this method are based on open loop hypothesis, which do not conform to the actual situation. In this paper, PBSIDopt (predictor-based subspace identification) class closed loop SID method, is applied to identifying a Linear Time-Invariant (LTI) model of a hovering tiltrotor UAV attitude dynamics. The method is compared with open loop SID. Result shows the model identified by PBSIDopt could matches the real flight data and have better fitting effect than that by open loop SID.