平台式导引头隔离度模型在线辨识方法

To solve the problem of online identification of platform seeker disturbance rejection rate models, an improved convolutional neural network based seeker disturbance rejection rate model identification method is proposed to realize efficient identification of the model generated by different torques and radome errors. Firstly, the seeker model of the platform is established, the disturbance rejection rate transfer function is derived, the guidance loop platform based on the parasitic loop is built, and the line of sight angle rate information under the disturbance of the missile body is obtained as the training and test data. Then, convolutional neural network is used to extract and reduce the feature dimension of the line of sight angle rate signal. Finally, the model diagnosis results are output by classification. The simulation results show that the proposed identification method of disturbance rejection rate model recognition accuracy can reach 99. 7 %, which improves the identification accuracy and fast online identification ability compared with traditional methods, and has a good engineering application prospect.