基于分形特征和ILST-KSVC的调制方式识别

Modulation Signal Identification Based on Fractal Theory and ILST-KSVC

针对传统的信号调制识别方式在信噪比较低的情况下识别精度低与种类少的问题,提出了一种新的基于分形理论及多分类最小二乘双支持向量机的通信信号识别方法.首先采集数字信号,对接收到的信号进行预处理,提取其分形特征作为识别的特征参数,然后采用多分类最小二乘双支持向量机分类器实现对未知信号的识别.该方法与传统的神经网络分类法及决策树分类法相比,具有更好的泛化推广能力.实验仿真结果表明,该方法在低信噪比情况下,调制识别准确率要优于其他调制识别方法,且在信噪比SNR>-5dB时,平均识别成功率达到91%以上.

Aiming at the problem that the traditional signal modulation identification method has low recognition accuracy and few identification types when the signal-to-noise ratio is low,a new communication signal recognition method based on fractal theory and least squares twin multi-class classification support vector machine(ILST-KSVC)is proposed.Firstly,the digital signal is collected.Then,the received signal is preprocessed.After that,its fractal feature is extracted as the feature parameter of the recognition.Finally,the ILST-KSVC classifier is used to recognize the unknown signal.Compared with the traditional neural network classification method and decision tree classification method,this method has better generalization and promotion ability.The experimental simulation results show that the digital signal modulation identification method proposed in this paper has better modulation recognition accuracy than other modulation identification methods at low signal-to-noise ratio.Moreover,when SNR>-5 dB is used,the average recognition success rate of this method reaches more than 91%.