引入合同变换矩阵的网络攻击信号盲分离算法

Blind Separation Algorithm of Network Attack Signal Based on Contract Transformation Matrix

对网络攻击信号进行盲分离,实现对攻击信号的准确有效检测。传统的网络攻击信号检测算法使用时频分析方法,提取非平稳群攻击信号的时频特征,实现信号检测,但算法把网络入侵检测正确率作为约束目标函数进行同步最优特征子集求解,复杂度较高,提出一种引入合同变换矩阵的网络攻击信号盲分离算法。采用时频分析Viterbi算法,得到信号谱的平均频率等于瞬时频率的时间平均,根据合同变换矩阵,对攻击信号进行离散数据解析化处理,构建网络攻击信号的解析模型,得到网络统计信号在多复变边界条件下的时频特征,实现盲分离算法改进。仿真实验表明,该算法能有效实现对网络攻击信号的盲分离,盲分离结果能准确反映网络攻击信号的内部特征,提高了对网络攻击信号的检测能力,对攻击信号的检测性能有所提高,保证了网络安全。

The blind separation of network attack signal, realize accurate and effective detection of the attack signal. Frequency analysis method using the network attack of traditional signal detection algorithm, extraction of non stationary signal time-frequency feature group attack, the realization of signal detection, but the algorithm to network intrusion detection correct rate as the objective function for synchronization constraints optimal feature subset solution, high complexity, algorithm a separation into the contract transformation matrix of the network attack signals blind is proposed. Using the time-frequency analysis of Viterbi algorithm, the time averaged mean frequency signal spectrum is equal to the instantaneous frequency,according to the contract transformation matrix, the attack on the signal analytical treatment of discrete data, analytical model construction of network attack signal, get the network statistical signal in multi complex time-frequency characteristics of boundary condition, to realize the blind separation algorithm is improved. Simulation results show that the algorithm can effectively realize the blind separation of network attack signals, blind source separation results can accurately reflect the internal characteristics of the network attack signals, improves the ability to detect attacks on the network signal, the detection performance of the improved signal to attack, to ensure network security.