基于通用化等效模型的飞机系统辨识算法

This study provides a frequency domain identification approach aimed at the high-order time-delay model matching problem for complex systems like carrier aircraft. A high-order time-delay model of the complex system is ultimately identified after the generalized frequency-domain identification model is established and the model is iteratively trained by system frequency domain response data using the gradient descent method of adaptive learning rate. This allows the model’s frequency domain parameters to automatically converge.Verification is carried out with the Dutch roll equivalent model of carrier aircraft. Through the identification and comparison of high-level and low-level models, the algorithm in this paper is compared with the traditional algorithm, and the frequency domain and time domain analysis show that the algorithm in this paper has a good identification effect. It solves the problem of direct fitting of high-order time-delay models, and is suitable for universal high-order time-delay models.