Anomaly Detection Under Multiplicative Noise Model Uncertainty

State estimators are crucial components of anomaly detectors that used to monitor cyber-physical systems. Many frequently-used state susceptible model risk as they rely critically on the availability an accurate state-space model. Modeling errors make it more difficult distinguish whether deviations from expected behavior due anomalies or simply a lack knowledge about system dynamics. In this research, we account for uncertainty through multiplicative noise framework. Specifically, propose use LQG based compensator in setting hedge against risk. The size residual estimator can then be compared threshold detect anomalies. Finally, proposed detector is validated using numerical simulations. Extension state-of-the-art detection systems handle represents main novel contribution present work.