Attack-resilient state estimation with intermittent data authentication

Network-based attacks on control systems may alter sensor data delivered to the controller, effectively causing degradation in performance. As a result, having access accurate state estimates, even presence of measurements, is critical importance. In this work, we analyze performance resilient estimators (RSEs) when any subset sensors be compromised by stealthy attacker . Specifically, consider with well-known l 0 -based RSE and two commonly used sound intrusion detectors (IDs). For linear time-invariant plants bounded noise, define notion perfect attackability (PA) result unbounded estimation errors while remaining undetected employed ID (i.e., stealthy). We derive necessary sufficient PA conditions, showing that system can perfectly attackable if plant stable. While prevented use standard cryptographic mechanisms (e.g., message authentication) ensure integrity under network-based attacks, their continuous imposes significant communication computational overhead. Consequently, also study impact intermittent authentication has guarantees attacks. show messages from some are intermittently authenticated, could not errors.