Signals Time Frequency Features Analysis And Classification For Civil Radar

The signal classification and recognition technology is one of the key problems in modern signal processing field.This technology is designed to classify and identify the unknown signal sources.Based on the theory of time-frequency analysis,this paper analyzes the characteristics of civil flight meteorological radar signals and thus uniquely identifies unknown signal,which provides prediction analysis under severe weather and rate of recognition for the key issues such as civil aircraft detection and civil airport management.The main work and research results of this paper include:Discuss on the major factors related to unintentional modulation of civil flight radar signals so that theoretically analyze the reason of radar signal feature.According to the characteristics of radar transmitter inner mechanism,this thesis summarize the influences on signal individual unintentional modulation feature caused by phase noise.What's more,simulation and implementation of realizing unintentional modulation refers to adding phase noise to three types of radar signal models including Single Frequency(SF)radar signals,Linear Frequency Modulation(LFM)radar signals,Phase Modulation(PM)radar signals.An ambiguity function mask(AF-mask)based on compressed sensing are proposed for radar feature.The feature utilizes the idea of compressed sensing,so that signals could be locally concentrated and optimized with sparsity.In this way,a more representitive time frequency feature in ambiguity domain with less data size will be extracted.Compared with ambiguity function representative slice,AF-mask covers much more time varying information of signals and possesses more accurate and more stable classfication rate in civil flight meteorological radar classification results.Meanwhile,AF-mask can realize signal reconstruction and instantaneous frequency.In addition,the experiment shows that AF-mask avoids high dimension redundancy because compressed sensing norm optimization can calculate the size of AF-mask.Feature optimization methods based on AF-mask for engineering are proposed.On one hand,we extract AF-mask feature according to the characteristics of signals.On the other hand,AF-mask feature can be extracted based on energy of signals.From the experiment, the two algorithms own a good accuracy and better run time although lower than compressed sensing mask.From an engineering point of view,the two fast optimization methods improve the efficiency and save the space of whole radar identification system.Create,collect and build big databases for signal identification.These big database including measured civil flight meteorological radar signals data and semi-physical simulation data provide a better evidence for experiments verification.Meanwhile,Radar Wave Adaptive Software based on matlab-GUI is designed for a better understanding and visualization on database and signal wave.