Instrumental variable covariance method and asymptotic analysis for the aircraft flutter model parameter identification

When the observed input-output datas are corrupted with observed noises in the aircraft flutter statistic model, we should obtain the more exact aircraft flutter model parameters to predict the flutter boundary accuracy and assure flight safety of plane. So we combine the instrumental variable identification method in system identification theory and covariance matching method in modern spectrum theory to get a new strategy-instrumental variable covariance method. In aircraft flutter´s statistic model, we introduce some instrumental variable to develop a covariance function. And we propose a new criterion function which is composed as a difference between the theory value and actual estimation value of the covariance function. Now the new criterion function based on the covariance function can be used to identify some unknown parameter vector in parameterization frequency domain response function, furthermore we give the procedure in detail to solve the new criterion function and correspond to the partial derivatives expression. By virtue of the accuracy analysis theory in system identification setup, we derive the asymptotic covariance matrix expression which is obtained from this paper´s instrumental variable covariance method. Then we can use this asymptotic covariance matrix expression to judge the effectiveness of this new identification method and design the external input excite signal. Finally we apply this new identification method to identify the transfer function in current loop of flight simulator and aircraft flutter model parameter identification. The simulation with real flight test data shows the efficiency of the algorithm.