Research Of Multi-node Information Acquisition In Communication Network Based On Interception Signal Analysis

Electronic reconnaissance refers to the use of electronic reconnaissance equipment to intercept the electromagnetic signals emitted by the target radar,communication station or other communication facilities,and obtain the target position information,strategic deployment or platform type information through technical means.Electronic reconnaissance is an important part of electronic countermeasures.Electronic reconnaissance is usually based on analysis of intercepted signals.In view of the diversity of intercepted signals,it is of great value to study different multi-target information acquisition technologies.In recent years,machine learning,deep learning and other related technologies have been increasingly applied in electronic reconnaissance,information processing and other fields.This paper focuses on the multi-node information acquisition technology based on machine learning.Among them,it is of great significance to study the technology of multi-objective data association and identification of interobjective connection.This paper focuses on the analysis of two kinds of intercepted signals(message information and on-off sequence),aiming at the problems in electronic reconnaissance,proposes the passive multi-station multi-target direction-finding data association algorithm based on clustering method and the node connection identification algorithm based on convolutional neural network,and verifies the effectiveness of the algorithm through simulation.The main work of this paper is as follows:1.Two different intercepted signals are analyzed: message message and on-off sequence.From the characteristics of the two kinds of intercepted signals,the differences of different signal applications are analyzed,which provides a theoretical basis for the subsequent research on data association and relationship identification.2.A passive multi-station multi-objective direction-finding data correlation method based on clustering method is studied.Firstly,based on message information,clustering technology is used to realize target attribute correlation and spatial correlation,and the performance of different clustering algorithms is analyzed by simulation.Then,the consistency of attribute correlation and the difference of spatial correlation in different clustering techniques are summarized.Finally,the effects of different observation errors on spatial correlation are compared.This method solves the difficulty of locating multiple false target points in the case of multi-station and multi-target in traditional passive positioning.3.This paper studies a method to identify the general relationship of nodes based on convolutional neural network.Firstly,based on the data of communication behavior,a graphical method of multi-node on-off sequence is proposed.Then,the normalized input of the model is obtained by means of image processing,so that the convolution neural network technology can be used to realize the recognition of the target connection relation.Then,the difference of recognition effect of convolutional neural network under different noise ratio is analyzed by simulation.Finally,the effects of different training sample Numbers,different batch training sample Numbers,different optimization methods and different pooling methods on the convolution neural network recognition are compared.The method is malleable and can be applied to different scenarios with different number of targets.