Interfere With Recognition Technology In The Communications Countermeasure Research

As the information era is coming, weapons and equipment becomes more electronic and informational, the modern warfare have gradually shifted the focus to electronic warfare and information warfare. As the unique combat methods and performance, communication countermeasure are an important part of electronic warfare. In the communication process, communication party interference signals can be identified if the type of interference can take the appropriate measures to avoid or suppress the maximum interference. Therefore, how to identify the type of the interference signal is the research focus of this article.The interference signals in communication countermeasure have many types and are defined differently, according to different classification rules. This article mainly research and analysis the four typical interference signal, such as the single tone interference, the multi-tone interference, the noise of part band interference and the sweep frequency interference. To detect the type of the interference signals is the problem of pattern recognition, related to the signal preprocessing, feature extraction and classification algorithms and other parts.For the aspect of the signal pre-processing, we use a linear function of the method of maximum and minimum to normalize the signals, and then use the method of Z score to centralize them, and Finally, use the digital filter to denoise the sample signal. By the normalization, centering, and smoothing of the filter can significantly improve the robu-stness of the data and can be manipulative.For the aspect of the signal's feature extraction, we use six parameters, namely, the carrier factor coefficients, the average flat spectrum coefficient, the kurtosis coefficien--t, the skewness coefficient, the R parameters and the gaussian white noise factor. After the simulation of the four interference signals, the results show that the six parameters are separable with good degrees.For the aspect of classification, the paper identify the four types of interference signals mainly based on BP neural networks, support vector machine and decision tree theory. The simulation analysis shows that the three algorithms have good recognition performance:when the jamming-noise ratio is greater than 5dB, the recognition rate is greater than 90%. The decision tree theory has better performance than BP neural networks and support vector machines when the jamming-noise ratio in the lower range. The performance of the latter two algorithm are the same basely, when the jamming-noise ratio is less than 5dB, the performance of recognition algorithm based on support vector machine is better than the recognition algorithm based on BP neural networks.