Abstract
The performance of the servo system has a great influence on aircraft attitude control. In application, servo system has nonlinear characteristics such as friction, clearance, zero position and so on which make it difficult to precise modeling. In engineering application, we often use the approximately linearized modeling method. However, it has its localization and cannot accurately describe the real condition of the servo devices. To test the correctness of the control algorithm in ground test, the real aircraft servo devices are often added to the whole closed-loop test system in Hardware-In-Loop-Simulation (HILS). In this paper, we introduce the traditional linearized modeling method used in HILS. To deal with the problem of nonlinear modeling, we propose a modeling method using BP neural network. Utilizing the large amount of data produced in HILS, a relatively precise model is trained. Compared with the linear modeling method, the effectiveness of neural network for nonlinear system modeling is verified. The presented method has some engineering application values.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call