Abstract:
Current frequency-domain system identification methods require an open-loop experiment design to identify the bare airframe dynamics of rotorcraft. Trained test pilots are needed to fly these open-loop flight tests on an unstable air vehicle and thus open-loop rotorcraft system identification is very challenging. In this paper, a novel closed-loop system identification approach is presented. This approach does not require pilot controls but applies Generalized Binary Noise as automatic test signals to all helicopter controls simultaneously. The bare airframe dynamics of the ACT/FHS rotorcraft simulator are identified by the optimized Predictor-Based Subspace Identification method in the time domain using these automatic maneuvers. Robust settings for the frequency range of interest of the chosen test signal and the overall experiment time are determined. Furthermore, the approach is evaluated in representative conditions including measurement noise and weak turbulence. Very promising results are obtained in comparison to classical frequency responses.