基于非平稳信号时频分析的网络攻击检测算法

NETWORK ATTACK DETECTION ALGORITHM BASED ON TIME-FREQUENCY ANALYSIS OF NON-STATIONARY SIGNAL

在复杂网络环境下,网络攻击特征信息通常表现为一组非平稳宽带信号,通过信号检测方法实现网络攻击检测,保证网络安全。传统方法采用傅里叶变换方法进行网络攻击的非平稳信号检测,由于傅里叶变换的时变性会引起较大的包络振荡,检测性能不好,提出一种基于非平稳信号时频分析的网络攻击检测算法。构建了复杂干扰环境下的网络攻击信号模型,提取网络攻击非平稳宽带信号的时频特征。采用WVD-Hough时频变换实现对网络攻击非平稳宽带信号的时频聚集,采用混叠谱模糊度函数分析频谱特征。得到网络攻击信号的瞬时频率估计结果,设计匹配滤波算法进行信号抗干扰设计,最后输出检测结果。仿真实验表明,采用该算法进行网络攻击检测,准确检测概率较高,检测性能优越。

In complex network environment, the network attack characteristic information is usually expressed as a set of non-stationary broadband signal ,which guarantees the network security through the signal detection method to achieve the network attack detection. In the traditional method, the Fourier transform method is used to detect the non-stationary signal of the network attack, and the detection performance is not good because of the large envelope oscillation caused by Fourier transform. Thus, a network attack detection algorithm based on the time-frequency analysis of non-stationary signal is proposed. A network attack signal model in complex interference environment is constructed to extract the time-frequen- cy characteristics when network attacks non stationary wideband signal, adopting the WVD-Hough transform frequency to realize the time-frequency gathering when network attack non stationary wideband signal and using aliasing ambiguity spectrum function to analyze spectrum feature to get the instantaneous frequency estimation results of network attack sig- nal and design match filtering algorithm for anti-jamming design of the signal and the final output test results. Simulation experiments show that the algorithm is used to detect the network attacks, the accuracy of the detection is higher and the detection performance is superior.