The ability to detect the linear attack of WL-CUSUM and FMA algorithms

The problem of detecting linear attacks on industrial systems is presented in this paper. object attacked by attack the wireless communication process from sensors to controller with simulated mathematical model (stochastic dynamical and random noises). matrices are calculated ensure that Kullback-Leiber (K-L) algorithm passed. With these matrices, window limited cumulative SUM (WL-CUSUM) finite moving average (FMA) utilized detect changes sequence residuals generated Kalman filter method appreciated ability attack. results show an appropriate range threshold WL-CUSUM FMA can be chosen case K-L cannot detect. Moreover, tested using Monte Carlo simulation also evaluation performance detection better than WL-CUSUM, CUSUM, Chi-squared (Chi2).