Design And Implementation Of Radar Type Recognition System Based On Machine Learning

The radar transmits electromagnetic wave to irradiate the target and receives its echo,so as to obtain the target's range,range change rate(radial velocity),azimuth and other information.With the development of radar,there are a series of electronic measures and actions to weaken radar reconnaissance effect by means of analysis,suppression and jamming,which are collectively referred to as radar electronic countermeasures.An important step of ECM is to identify the specific type of enemy radar according to the radar parameter information obtained by reconnaissance,which provides a solid foundation for further judging radar working mode,system platform,formation and other information.The existing recognition algorithm calculates the membership degree of several characteristic values of radar signal respectively,and finally obtains the overall membership value,and judges the type attribution of the signal according to this value.In view of the situation that the existing methods can not use historical data to participate in the recognition,and the recognition accuracy of complex signals is low,this paper proposes an artificial intelligence algorithm based on neural network and decision tree for radar type recognition new method.In this paper,a neural network is constructed to identify the types of radar signals successfully.In the case of missing and wrong test data,the signal recognition can still be completed correctly with high accuracy.By using decision tree algorithm to judge the threat level of unknown signal,the unified management of multiple characteristic parameters is completed,the coding process is simplified and the judgment accuracy is improved.Finally,this paper constructs a complete signal recognition software system,which is mainly divided into data acquisition,data cleaning,neural network construction,model identification,threat level analysis of unknown signals and so on.The requirements of each software module are decomposed,and the functional modules of the software are analyzed and designed in detail according to the requirements.This paper presents the programming implementation of the system and the input and output data structure of each module,and generates simulation data to analyze and verify the function and efficiency of the software.The software uses Python as the main programming language,and uses pyqt,numpy,Matplotlib and other software libraries and modules.The software has been used to analyze and evaluate the experimental data.At the same time,through the comparison with the existing engineering software,it is proved that the new method can provide better recognition accuracy and recognition speed.It plays an active role in radar type identification...