科技通报
科技通報
과기통보
BULLETIN OF SCIENCE AND TECHNOLOGY
2015年
4期
118-120
,共3页
污点数据%网络安全%攻击信号%虚警门限
汙點數據%網絡安全%攻擊信號%虛警門限
오점수거%망락안전%공격신호%허경문한
tainted data%network security%attack signal%false alarm threshold
对网络攻击信号的线性调频建模和虚警门限预估计，是实现网络攻击信号检测的基础。传统方法采用路由数据交换的阵列信号聚类方法进行信号建模和虚警门限估计，由于攻击信号具有多波束自相似特征，导致虚警门限的预测精度不高。提出一种基于污点数据双模聚类的攻击调频信号虚警预估计算法，采用频率调制规律设计网络攻击信号模型，对攻击信号进行双线性核相位加权，得到脉冲压缩后的攻击信号二次调频信号输出。在双模聚类的时频平面内实现对信号检测能量的聚集，以改善信号分量聚集的尖峰，实现了对攻击调频信号的时间点与频率点的重组排列和特征分布估计，达到对虚警门限准确估计的目的。实验结果表明，该算法能有效估计攻击调频信号的虚警门限，对网络攻击信号的虚警门限的预估计精度达到98.5%，提高了对攻击信号的参数估计和信号检测能力，提高检测概率，优越性明显。
對網絡攻擊信號的線性調頻建模和虛警門限預估計，是實現網絡攻擊信號檢測的基礎。傳統方法採用路由數據交換的陣列信號聚類方法進行信號建模和虛警門限估計，由于攻擊信號具有多波束自相似特徵，導緻虛警門限的預測精度不高。提齣一種基于汙點數據雙模聚類的攻擊調頻信號虛警預估計算法，採用頻率調製規律設計網絡攻擊信號模型，對攻擊信號進行雙線性覈相位加權，得到脈遲壓縮後的攻擊信號二次調頻信號輸齣。在雙模聚類的時頻平麵內實現對信號檢測能量的聚集，以改善信號分量聚集的尖峰，實現瞭對攻擊調頻信號的時間點與頻率點的重組排列和特徵分佈估計，達到對虛警門限準確估計的目的。實驗結果錶明，該算法能有效估計攻擊調頻信號的虛警門限，對網絡攻擊信號的虛警門限的預估計精度達到98.5%，提高瞭對攻擊信號的參數估計和信號檢測能力，提高檢測概率，優越性明顯。
대망락공격신호적선성조빈건모화허경문한예고계，시실현망락공격신호검측적기출。전통방법채용로유수거교환적진렬신호취류방법진행신호건모화허경문한고계，유우공격신호구유다파속자상사특정，도치허경문한적예측정도불고。제출일충기우오점수거쌍모취류적공격조빈신호허경예고계산법，채용빈솔조제규률설계망락공격신호모형，대공격신호진행쌍선성핵상위가권，득도맥충압축후적공격신호이차조빈신호수출。재쌍모취류적시빈평면내실현대신호검측능량적취집，이개선신호분량취집적첨봉，실현료대공격조빈신호적시간점여빈솔점적중조배렬화특정분포고계，체도대허경문한준학고계적목적。실험결과표명，해산법능유효고계공격조빈신호적허경문한，대망락공격신호적허경문한적예고계정도체도98.5%，제고료대공격신호적삼수고계화신호검측능력，제고검측개솔，우월성명현。
The linear frequency modulation model and false alarm threshold estimation of the network attacks signal are the foundation to realize the network attack detection. The traditional methods use array signal clustering routing for data ex?change signal modeling and false alarm threshold estimation, because the attack signal has multi beam and self similar characteristic, the precision of false alarm threshold is bad. A false alarm pre estimation algorithm of attack FM signal is proposed based on tainted data dual clustering, network attack signal model is designed with frequency modulation princi?ple, bilinear nuclear phase weighting is taken for the signal. The two FM signal output is obtained after pulse signal pro?cess, the signal detection energy is gathered in the time-frequency plane with dual bimodal clustering. The signal compo?nent accumulation peak is improved. Accurate estimation of false alarm threshold is achieved. The experiment results show that the algorithm can effectively estimate the attack frequency modulated signal alarm threshold, estimation accuracy is 98.5%, the parameter estimation and signal detection capability of the signal are improved, the detection probability is im?proved. It has obvious superiority.
