雷达信号调制方式识别与参数估计算法研究-通信与信息系统专业论文.docxVIP

AbstractFighting Abstract Fighting in modem military electronic reconnaissance and electronic countermeasure，through to the identification and parameter estimation of the intercepted radar signal source，and on the basis of the results and specific parameter values，according to certain rules and algorithm and querying the emitter characteristics database，we can determine the purpose or type of radar signal from which enemy platform．After that，the reference for we make operational plans is provided，SO we can do a start system and obtain the initiative of the battlefield． In recent years，in the emergence of a lot of radar signal identification algorithm， neural network theory and intra-pulse analysis algorithm have been widely applied，but they each have their own shortages．And a single method cant meet the increasingly complex modulation mode signal source identification and parameter estimation．In this paper,the combining the neural network theory and intra-pulse analysis is applied to radar signal identification and parameter estimation．And the emphasis is how to use the self-learning ability of neural network to ascend with the serious distortion parameters，disturbance pulse string of radar signal identification accuracy,solve some of the deficiencies and defects in analysis method．In this paper,the main work and some contributions are： (1)First，it introduces the real signal of plural representation and deduces the classic complete process formula of SNR estimation algorithm，and introduces in Cramer-Rao inequality by which the most theoretical value of performance comparison is given．Then，it also introduces two important related technology in the process of signal processing，that is curve fitting and smoothing filtering．What’S more，the simple introduction of several kinds of window function is given．At last,based on that some simulation and analysis are made to show the enhancement effect of performance and precision by smoothing filtering in the processing of signa