Research On Classification Of Airplane Targets Based JEM Signal

With the rapid progress of modern technology, modern warfare has beentransformed into information warfare, therefore, we want to seize the initiative inmodern warfare, the key is to seize information superiority. Radar accomplishes a way ofaccess to information superiority and it also plays an important role. Improvement anddevelopment of modern radars are the prerequisites for the emergence and developmentof radar target recognition. As the narrow-band radar target recognition improvesconstantly, it will provide a new target recognition approach. At the same time, digitaltechnology also contributes to the rapid development of radar target recognition fromtheory into the project. This dissertation mainly focuses on research on classification ofairplane targets based on JEM (jet engine modulation) signal. The main contents are asfollows:Firstly, we introduce the research background and significance, and describe theecho waveform based on the direct extraction methods and others commonly appliednarrow-band radar target recognition methods.We outline the study of the simulation data of three types of aircraft, introduce thegenerating mechanism of modulation features and the ideal parameter model of rotor ofJet Engine Modulation (JEM),analyze the impaction of the actual aircraft toward theideal parameter model, and study the impaction of some parameters such as pulserepetition frequency and dwell time toward JEM modulation feature based on the pointsmentioned.We also detail the feature extraction and classification algorithms of the three typesof aircraft target of narrow-band radar, briefly overview the pattern recognitiontechnology and its application in radar target recognition, specifically describe the threetypes of aircraft’s motion characteristics and feature extraction method of themodulation cycle of rotating parts, generally summerize SVM (support vector machine)classification of the three types of aircraft targets, which includes the theory of machinelearning, the principle supporting the SVM algorithm and SVM’s multi-classclassification, and study the classification of the three types of aircraft based onmeasured data from the analysis of three aspects of measured data, two-dimensionalcharacteristics and identification results of the three types of aircraft.We then present the narrow-band radar echo simulation and automatic targetclassification aided design software, which including four modules such as designing background and significance, Matlab GUI programming introduction, the softwarefeatures and system functions and software manual. At the same time, software manualwhich covers eight parts of the software installation, software operation, loading testdata, generating simulation data, data features management, object recognitionclassification, data backup and data recovery.Finally, the summary of the dissertation is made and the problems in furtherresearch are indicated.