Unscented Kalman filtering for nonlinear systems with sensor saturation and randomly occurring false data injection attacks

Abstract

In this paper, an unscented Kalman filtering problem is studied for a nonlinear system with sensor saturation and randomly occurring false data injection attacks. A random variable obeying the Bernoulli distribution is employed to characterize the phenomena of the randomly occurring false data injection attacks. The aim of this paper is to design a modified unscented Kalman filter by minimizing an upper bound of filtering error covariance. Furthermore, a sufficient condition is provided to ensure an exponentially bounded filtering error in the mean square sense. Numerical simulations are presented to illustrate the validity of the proposed filter.

Model‐based fault diagnosis of networked systems: A survey

“…H∞ filtering and LMI Fading measurements [11,23] Quantization effects [74] Stochastic additive faults [15,28] Medium access constraints [14] Unknown transition probability [75] Comprehensive incomplete measurements [12,13,25,26] EM algorithm under the Bayesian framework Asynchronous measurements in distributed systems [54] Residual generating based on fault diagnosis filters and observers Additive faults & incomplete measurements [7,8,17,24,27,31,32,46,47] Attacks on sensors [73] Soft faults & packet dropouts [48] Actuator faults [42,45] Faulty periodic communication [30] Cyber attacks [39,76] Sliding mode observer Attacks on sensors [72] Unknown input observer False data injection attacks [35] Minimum-variance filtering and Kalman filtering Cyber attacks [20,37,64,70] Additive faults [49] Homomorphic encryption [77] Particle filtering Cyber attacks [41,69] Strong tracking filtering Packet dropouts [78] Distributed resilient filtering Sensor degradation [79] Self-learning approaches Additive sensor fault [71,75,…”

Secure state estimation for unmanned aircraft cyber‐physical systems under multiple attacks

“…In the 2015 Ukraine Blackout, attackers were able to wrest control from substation control centers through false data-injection attack [27]. Referenced by previous studies [2][3][4][5], the data-injection attacks for controller and sensor are modeled as dual unknown interference inputs to deal with any possible attacks. Through designing the effective method, the dual interference decouple can be decoupled in the filtering process without requiring any known attack information.…”

“…The attacker may implement DoS attack on the sensor output communication link, resulting in the loss of transmission data. Besides, studies in Lu et al [4] and Mo et al [5] show that a new type of attack called a false data-injection attack can be directed at the system if some system parameters are known to the attacker. The attacker may maliciously inject false signals into the communication link from controller to actuator and the sensor to observer.…”

Deception attack monitoring in vulnerable hydroelectric generator system

“…In Wang and Xu [14], a guaranteed cost control problem for a CPS under Dos jamming clear periodic attack with inherent packet dropout is solved via both state feedback and output feedback methods. In previous studies [15][16][17], a Kalman filtering problem is studied for nonlinear system with sensor saturation and randomly occurring false data injection attack. In Geng and Liu [18], a stability guaranteed self-triggered scheme is proposed to actively defend the denial of service (Dos) attacks.…”