基于软K段主曲线的信号细微特征识别

IDENTIFYING IMPERCEPTIBLE FEATURE OF SIGNALS BASED ON SOFT K-SEGMENTS ALGORITHM FOR PRINCIPAL CURVES

为解决同型号电台信号识别问题,针对具有同步头的信号,根据噪声在频域分布于底层的特点,从频域角度出发,建立基于软K段主曲线算法的功率谱骨架模型。将谱骨架作为暂态特征,计算其信息维数和盒维数,将得到的分形维数使用SVM分类器进行训练,实现对不同电台发送的相同数据波形信号的识别。实验结果表明,在信噪比较低的情况下,该方法可以有效识别电台,识别率仍可达到80%以上。

In order to solve the problem of signals identification among transmitters with same model, in the paper, aiming at the signals with synchronous head and according to the characteristics of noise distributing on underlayer in frequency domain, we build a power spectrum skeleton model proceeding from the perspective of frequency domain, which is based on the soft K-segments algorithm for principal curves. We take the skeleton as transient feature and calculate its information dimensions and box dimensions, the derived fractal dimensions are trained by SVM classifier, the identification of waveform signals with same data sent by different transmitters is then implemented. Experiment results show that under the condition of low SNR the method can effectively identify the transmitters, the recognition rate can still reach 80 percent or higher.