Resilience and Performance Analysis for State Estimation against Integrity Attacks

We consider the problem of resilient dynamical state estimation in the presence of integrity attacks. We conduct resilience and performance analysis for a convex optimization based estimator. The employed approach for analyzing resilience of an estimator is novel and generic for a wide class of estimators and thus can achieve greater generality, as long as an estimator can be decomposed into a convex optimization based form. We show sufficient and necessary conditions for resilience with a trivial gap. The tradeoff between minimum mean square error (MMSE) optimality and resilience is well investigated. Due to the constructive proof of the main results using the force analogy, we present an upper bound on the damage an attacker can cause when the sufficient condition is satisfied. Simulation results are also given to validate the resilience and performance analysis.