Research On Frequency Hopping Signal Detection Methods In Complex Electromagnetic Environment

The frequency hopping communication system has the advantages of strong antiinterference performance,strong anti-interception ability,and good anti-fading effect.It has a wide range of applications in military and civilian communications.On the battlefield,radio communication equipment and electronic reconnaissance equipment are constantly sending electromagnetic waves into space.At the same time,more and more communication signal patterns are used,resulting in a very complicated electromagnetic environment in the communication process.The prerequisite for obtaining the enemy's frequency hopping communication information is to detect the intercepted signal,which provides a basis for further analysis and processing of the frequency hopping signal.Therefore,it is of great significance to study the detection method of the frequency hopping signal in a complex electromagnetic environment.The frequency hopping signal is a time-varying signal.Its frequency changes continuously with time.It is not easy to describe the characteristics of the signal with precise mathematical expressions.However,the time-frequency image of the frequency hopping signal shows the change of the signal.The rules and features can be used to complete the detection of frequency hopping signals through time-frequency image processing.This article studies the time-frequency image features and deep learning frequency hopping signal detection algorithms.The main content is divided into the following aspects:1.Research on frequency hopping signal detection algorithm based on image features.The electromagnetic environment of the frequency hopping signal is analyzed,and the detection algorithm research is carried out according to the image characteristics of the interference signal and the frequency hopping signal.Because noise has a greater impact on image processing,morphological operations are used to remove the interference of irregular noise;for the problem of image feature widening after morphological operations,image thinning operations can effectively reduce the amount of image feature data,thereby reducing Difficulty of image recognition;Hough transform has the characteristics of detecting objects of specific shapes through voting methods,used for image feature extraction,and combined with the characteristics of frequency hopping signals to complete the signal detection;finally set up a software development kit GNU Radio and hardware USRP The combined software radio platform is designed and collected through the physical platform and the measured frequency hopping signal data,and the feasibility of the algorithm is verified through the analysis and processing of the data.2.The use of deep learning for frequency hopping signal detection is analyzed and discussed.In order to be able to quickly and accurately detect frequency hopping signals under low signal-to-noise ratio,a convolutional neural network(Convolutional Neural Network,CNN)method is used for signal detection research.Convolutional neural network has the feature of automatically extracting image features.It can be used to process the time-frequency diagram of frequency hopping signals,design and analyze the network structure and perform parameter optimization experiments,and then obtain the optimal network parameters and signal detection model,which is verified by simulation experiments.The method can quickly and accurately detect frequency hopping signals under low signal-to-noise ratio,and at the same time improve the detection and recognition rate.In order to complete the frequency hopping signal detection and signal marking,the Faster RCNN method is used for frequency hopping signal detection.The target detection algorithm has the ability to identify and mark multiple targets.It can be used to process frequency hopping time-frequency diagrams in complex electromagnetic environments.The method is applied to the field of frequency hopping signal detection,the network model is trained and tested,and the feasibility of the algorithm is verified through simulation experiments.