针对复杂电磁环境下正交跳频信号分选时效不高问题,提出了一种正交跳频信号动态分选方法。首先基于滑动窗口的数据流模型,采用构造型神经网络对跳频信号频域数据和方位信息动态聚类,减轻噪声等因素影响,解决方位信息和幅度关联模糊性的问题;再在相应的覆盖簇内运用时频关联方法,实现正交跳频信号的动态分选,实验结果表明该方法是有效的。
In this paper, a new method for dynamically sorting orthogonal FH (Frequency Hopping ) signals is presented to solve the problem of sorting the orthogonal FH signals in the complex electromagnetic environment. The method based on the sliding window' s data stream models takes advantage of the constructive neural network to cluster frequency- domain data and orientation information dynamically, it reduces the influence of noises and solves the correlation fuzziness of the orientation information and the amplitude. The orthogonal FH signals are dynamically sorted by using time-frequency correlation method in each covered cluster. Experimental results have shown that the method is efficient.