Dynamic identification of key nodes in digital microwave communication link based on network topology

Abstract

In order to solve the problems of the traditional microwave communication node identification methods, such as the large difference between the number of identification and the actual situation, and the low accuracy of identification, the dynamic identification of key nodes of digital microwave communication link based on network topology is studied. The data labels of key nodes are set by supervised machine learning, and the synchronous signals of four key nodes close to the digital microwave emitter are selected to prevent the collected signals from excessive noise. Fuzzy clustering is used to process the collected recognition samples, and then the neural network model is trained to identify the key nodes. In order to prevent over fitting during training, L4 regularization method is used to add dropout to the hidden node neurons in each network topology. Then, the identification neural network model of network topology is used to train and dynamically identify different key nodes of the link to realize the dynamic identification of key nodes of digital microwave communication link. The experimental results show that the recognition accuracy of key nodes is 95%, and the recognition performance of key nodes is improved.