CAJ | 학술논문

为提高雷达辐射源信号识别速度，提出了一种脉内特征参数选择的新方法。首先对熵、相像系数、模糊函数、复杂度和双谱五种脉内特征参数的提取算法进行了研究，其次利用留一法误差来评估特征参数对分类器泛化性能的影响，最后对特征参数的评估进行了仿真实验。仿真结果表明，基于留一法误差可以较好地实现特征参数选择，为提高雷达辐射源信号识别准确率提供参考。
위제고뢰체복사원신호식별속도，제출료일충맥내특정삼수선택적신방법。수선대적、상상계수、모호함수、복잡도화쌍보오충맥내특정삼수적제취산법진행료연구，기차이용류일법오차래평고특정삼수대분류기범화성능적영향，최후대특정삼수적평고진행료방진실험。방진결과표명，기우류일법오차가이교호지실현특정삼수선택，위제고뢰체복사원신호식별준학솔제공삼고。
In order to improve the signal recognition speed of radar emitter, this paper gives a new method for selection of signal characteristic parameters in-pulse. In this scheme, extraction algorithms of five sorts of parame-ters in pulse such as the entropy, resemblance coefficient, fuzzy function, complexity and bi-spectrum, are studied at first, next, the impacts of characteristic parameters on generalization performance of classifier are evaluated by using the error rate of LOO and the evaluation of characteristic parameters is simulated. Simulation results show that the selection of characteristic parameters could be implemented well based on error of LOO, which offers a reference for improving the accuracy rate of signal recognition of radar emitter.