CAJ | 학술논문

网络入侵信号表现为一种多频非线性微弱信号，传统的随机共振检测方法受弱信号幅度和临界阈值约束，检测性能受限。提出一种采用随机共振概率密度特征确定信号检测临界阈值的方法，建立网络入侵信号自适应随机共振检测系统，把多频网络入侵信号采用EMD分解的方法分解为高频和低频信号两部分，使每个频段满足随机共振条件。采用相轨迹过零周期数相变判别算法提取随机共振概率密度特征，确定检测系统的临界阈值，实现对入侵信号的准确判别和检测。仿真实验表明，采用该算法进行网络入侵信号检测，能有效检测出多频非线性微弱入侵信号，检测概率具有明显提高，漏检和误报概率降低。
망락입침신호표현위일충다빈비선성미약신호，전통적수궤공진검측방법수약신호폭도화림계역치약속，검측성능수한。제출일충채용수궤공진개솔밀도특정학정신호검측림계역치적방법，건립망락입침신호자괄응수궤공진검측계통，파다빈망락입침신호채용EMD분해적방법분해위고빈화저빈신호량부분，사매개빈단만족수궤공진조건。채용상궤적과령주기수상변판별산법제취수궤공진개솔밀도특정，학정검측계통적림계역치，실현대입침신호적준학판별화검측。방진실험표명，채용해산법진행망락입침신호검측，능유효검측출다빈비선성미약입침신호，검측개솔구유명현제고，루검화오보개솔강저。
The network intrusion signal is a kind of multi frequency weak signal detection, and the stochastic resonance is traditional methods of detecting weak signal amplitude and threshold constraint. A method of determining the critical threshold signal detection method based on stochastic resonance probability density characteristics is proposed, the net-work intrusion signal adaptive stochastic resonance detection system is established. The method using EMD decomposition into two parts:high frequency and low frequency signal multi frequency network intrusion signal, and the phase trajectory zero cycle phase discrimination algorithm to extract the stochastic resonance probability density characteristics. The critical threshold detection system is determined, the intrusion signal recognition and detection is realized. Simulation results show that it can effectively detect the multi frequency weak signal, and detection probability is improved markedly with lower missing detection and false alarm probabilities.