Deception Attacks Research Articles

Performance analysis for cyber–physical systems under two types of stealthy deception attacks

This paper focuses on the performance analysis of cyber–physical systems under the stealthy deception attack. Two types of attack models, strictly stealthy attack and ϵ-stealthy attack, are considered. The strictly stealthy deception attack model is firstly formulated based on the difference equations of the normal and compromised systems. We then present a necessary and sufficient condition for the existence of the strictly stealthy deception attack. The design method for the attacker to achieve the strictly stealthy deception attack is also provided. Considering that the strictly stealthy attack strategy is associated with the undetectable points of the difference equations, we propose a computationally efficient algorithm to find such points. It is shown that these points form a subspace which can be obtained in finite recursion steps. The corresponding defense strategy is further designed by avoiding the initial state falling into the subspace. For ϵ-stealthy attack, a necessary condition is provided for its existence. Furthermore, due to the fact that the difference system matrix does not have an unstable eigenvalue, it is proved that the ϵ-stealthy attack does not exist. Simulation examples are presented to illustrate the theoretical results.

Fuzzy-based adaptive event-triggered control for nonlinear cyber-physical systems against deception attacks via a single parameter learning method

This paper investigates the adaptive fuzzy control issue of cyber-physical systems (CPSs) against unknown deception attacks and external disturbance. Based on the backstepping design framework, the uncertainty term composed of the approximation error term generated by the fuzzy logic system, the external disturbance and the deception attack signals are regarded as a whole in the recursive design process. Then, by combining the single parameter learning method with the adaptive fuzzy technique, the uncertain terms containing approximation error, disturbance and injection attacks can be transformed into the form of linear parameterization with only one unknown scalar parameter, which greatly reduce the calculation process. In addition, the dynamic event-triggered adaptive control technology can be integrated into the control design to reduce the amount of data transmission. Theoretical analysis shows that all the signals in the closed-loop system are bounded under the proposed control scheme, and the Zeno behavior is excluded. Finally, the two-stage chemical reactor and mass-spring-damper systems are employed to demonstrate the validity of the proposed adaptive fuzzy control solution.

Novel adaptive filter design for complex‐valued network systems under hybrid cyber attacks

AbstractThis paper is concerned with the problem of designing adaptive event‐triggered filter for complex‐valued network system under complex‐valued hybrid cyber attack. Firstly, the main contribution of the article is to propose a complex‐valued adaptive event triggered strategy to improve the resource utilisation in the network channel. Secondly, during the modeling process of the filtering error model, we introduce complex‐valued stochastic deception attacks and replay attacks that follow the Bernoulli distribution, thereby forming the complex‐valued hybrid network attacks investigated in this paper. Sufficient conditions for the asymptotic stability of the designed complex‐valued filtered error system are derived using Lyapunov stability theory. In addition, the parameters of the filter are derived using the linear matrix inequality (LMI) method. Finally, a numerical simulation example is given to verify the feasibility of the proposed method.

Output regulation for switched systems via dual distributed event-triggered mechanisms under multiple deception attacks

The secure output regulation (OR) is studied for networked switched systems (NSSs) that exhibit severely unstable dynamics (SUDs) and are vulnerable to multiple deception attacks. The occurrence of some destabilizing switching and unsolvable OR of each individual subsystem is permitted. Dual distributed event-triggered mechanisms (DDETMs) are proposed, incorporating both distributed event detection conditions to tackle the issue of unreliable single-channel data during deception attacks on sensor networks, and mode-matching detection conditions to reduce asynchronous intervals. A dwell time switching rule is developed without consistency constraint, i.e., the sub-controller's stabilizing/destabilizing switching may be inconsistent with that of the subsystem. Judging conditions are introduced for the occurrence of multiple deception attacks on dual-terminal sensor/transmission networks, which adjust the increment coefficients of Lyapunov functions before or after switching instants. Sufficient conditions are provided to ensure secure OR for NSSs under multiple deception attacks. With the aid of a switched RLC circuit, the effectiveness of the presented methodologies is validated.

Hybrid-triggered-based fault-tolerant dissipative control for T-S fuzzy parabolic PDE systems with external disturbances and deception attacks

This paper addresses the fault-tolerant hybrid-triggered control problem for a class of nonlinear parabolic PDE systems with Neumann-type boundary conditions subject to actuator faults, external disturbances and randomly occurring deception attacks. Precisely, in order to provide an accurate representation of the considered nonlinear parabolic PDE system, a Takagi–Sugeno fuzzy PDE model is utilized. Further, in an effort to reduce the amount of communication transmissions, a more general hybrid-triggering method is implemented which includes both time- and event-triggered scheme. For the descriptions of the hybrid-triggered strategy and deception attacks, stochastic variables that follows Bernoulli distribution are presented. At the same time, an actuator fault model is considered in order to manage the impact of faults that occur in the actuators of the system under consideration. By the construction of suitable Lyapunov-Krasovskii functional, a set of adequate requirements is procured that ensure the closed-loop system is asymptotically stable and dissipative. In addition, the necessary controller gain matrices are obtained by making use of linear matrix inequalities. In the end, two numerical examples are exhibited to articulate the efficacy of the proposed control design technique.

Resilient and event-triggered control of singular Markov jump systems against cyber attacks

This paper is concerned with the H ∞ resilient and event-triggered control of singular Markov jump systems against deception attacks. The deception attacks are modelled as a random bounded signal which is governed by a Bernoulli distributed random variable. The event-triggered scheme is adopted to achieve a trade-off between system performance and network resources. Based on the technique of stochastic Lyapunov–Krasovskii functionals and linear matrix inequalities, efficient criteria are developed such that the closed-loop system is stochastically admissible with a certain H ∞ performance under deception attacks. Then, the co-design of resilient controller gains and event-triggered rules is provided in terms of a group of feasible LMIs. Finally, two examples are employed to verify the validity of our design.

Dynamic event-triggered output feedback control for networked Markovian jump systems under hybrid attacks

This paper investigates the dynamic event-triggered control problem for a networked Markovian jump system under hybrid attacks. In the concerned system, both the sensor-controller (S-C) network and the controller-actuator (C-A) network suffer from hybrid attacks, including denial-of-service (DoS) attacks and deception attacks. Two dynamic event-triggers at the controller and the sensor are designed to handle these dual-network hybrid attacks. Moreover, a unified framework is constructed to design an observer-based output feedback controller (OBOFC) which simultaneously considers hybrid attacks and dynamic event-triggering strategies (DETSs). Sufficient conditions to ensure the input-to-state stability (ISS) of the concerned system are derived. Furthermore, a set of linear matrix inequalities (LMIs) are proposed to obtain the desired observer-based feedback control gains and event-triggering parameters. A simulation example is given to further verify the effectiveness of the proposed control and event-triggered schemes.

Towards invariant extended Kalman filter-based resilient distributed state estimation for moving robots over mobile sensor networks under deception attacks

This paper studies the resilient distributed state estimation problem for moving robots over wireless sensor networks (WSNs) under deception attacks. The state of a moving robot consists of its position and orientation, evolving according to a nonlinear motion model, and the WSNs are mobile. These settings make the problem more challenging than most existing researches on the resilient distributed state estimation which recover stationary or linear states by static WSNs. To achieve the resilient distributed state estimation for the robot over a mobile sensor network (MSN), we propose a novel invariant extended Kalman filter-based resilient distributed state estimation approach. Specifically, based on the mappings between a Lie group and its Lie algebra, we first define a new type of estimation errors and linearize it. Utilizing the linearization results, the calculation of the current prior estimate based on the last posterior one is achieved by each sensor, which avoids the estimation biases induced by the computation of Jacobian matrices. Then, sensors in the MSN fuse the local prior estimates of their neighbors according to the covariance intersection fusion rule, obtaining more accurate estimates than their local prior ones. Finally, measurements collected by the sensors are used to correct the estimates. In the correction process, an adaptive gain is designed for each sensor to restrict the negative effects of the attacks on the estimates, which is crucial for the accomplishment of the resilient estimation task. We illustrate the performance of the approach by convergence analysis of the estimation errors based on the boundedness of the covariance matrix estimates computed in the state estimation process and verify it by simulation and experimental results.

Observer‐based event‐triggered security control for nonlinear networked control systems with cyber attacks

SummaryThis paper investigates the observer‐based security control problem for interval type‐2 nonlinear networked control systems with deception attacks and replay attacks. Firstly, to alleviate the communication burden of the network, a novel adaptive event‐triggered communication (ETC) mechanism is proposed. Compared with existing observer‐based adaptive ETC security control techniques, the threshold of the proposed adaptive ETC scheme can be adjusted dynamically depending on the value of adjustable parameters and estimation error. In particular, the adjustable parameters can change the sensitivity of the threshold flexibly. Then, taking the influence of the adaptive ETC mechanism and cyber attacks into consideration, an observer‐based security controller is developed under the co‐designed method. Furthermore, by utilizing the Lyapunov stability theory, sufficient conditions for asymptotically mean‐square stable of the closed‐loop system are obtained. Ultimately, the feasibility of the designed control approach can be proved via a simulation example.

Event‐triggered networked cascade control systems design subject to hybrid attacks

This paper deals with the design problem of networked cascade control systems (NCCSs) with two different event‐triggered mechanisms (ETMs) for the feedback loops. A combination of stochastic denial‐of‐service (DoS) attack and deception attack has been considered for the communication network. To comply with the practical point of view, the primary and the secondary systems have been considered with distinct disturbances. First, the closed‐loop system with state feedback controllers is provided in a unified stochastic delayed system. Then, the Lyapunov–Krasovskii stability theory and stochastic analysis techniques are employed to derive and formulate control design conditions in terms of linear matrix inequalities (LMIs). Finally, a superheater steam temperature control is utilized to assess the applicability of the proposed method.

Extended dissipative performance for a complex dynamical networks based on adaptive event‐triggered filter method subject to hybrid attacks

SummaryThe challenge of adaptive event‐triggered based filter design for complex dynamical networks with time‐varying coupling delays under hybrid attack is studied in this research. A new model of complex dynamical networks with input constraints, involving denial‐of‐service (DoS) attack, replay attack, and deception attack is designed. An adaptive event‐triggered scheme is applied to filter design, which adaptively modifies the communication threshold level in order to reduce the effect of the transmission load. By using Lyapunov theory and inequality techniques, a novel necessary condition is proposed that guarantees mean‐square asymptotically stability with extended dissipative performance. The proposed filter parameters are then given an explicit expression, which is solved using a set of linear matrix inequalities. Finally, two numerical results are presented to demonstrate the validity of the theoretical result suggested.

Bipartite secure synchronization for dynamic networks under deception attacks via delay-dependent impulsive control

In this paper, we investigate mean-square bounded bipartite synchronization of coupled dynamic networks subject to deception attacks, especially under directed/undirected signed interconnection structures. An impulsive secure control input involving stabilizing delays is proposed. The average impulsive delay (AID) method is used to fetch the information of time delay existing in impulses, and then this information is effectively integrated to mitigate the negative effects of deception attacks on synchronization. A concept of stochastic variables is introduced to serve as a measure of attack success when delayed impulsive control signals are replaced with false data from the impulsive controller-to-actuator. To achieve desired synchronization performance under deception attacks, the gains of pinning impulsive control, the average impulsive interval (AII), AID and the attack intensity are co-designed. An upper bound for synchronization errors is provided, which can be adjusted with multiple adjustable parameters. In addition, the AII and AID methods are used to remove the restriction on upper/lower bounds of two consecutive impulsive instants. Finally, numerical examples are given to illustrate the validity of the theoretical results.

Predefined-time consensus tracking of high-order multiagent system with deception attack

Predefined-time consensus tracking of high-order multiagent system with deception attack

Sliding Mode Control for Sampled-Data Systems Subject to Deception Attacks: Handling Randomly Perturbed Sampling Periods.

In this article, the sliding mode control problem is addressed for a class of sampled-data systems subject to deception attacks. The sampling periods undergo component-wise random perturbations that are governed by a Markovian chain. The component of the sampled output is transmitted via an individual communication channel that is vulnerable to deception attacks, and Bernoulli-distributed stochastic variables are utilized to characterize the random occurrence of the deception attacks initiated by the adversaries. A sliding mode controller is designed to drive the state into the sliding domain around the specified sliding surface, and sufficient conditions are derived to guarantee the exponentially ultimate boundedness of the resultant closed-loop system in the mean-square sense. Furthermore, an optimization problem is established to pursue locally optimal control performance. Finally, a simulation example is given to verify the effectiveness and advantages of the developed controller design approach.

Concurrent Cyber Deception Attack Detection of Consensus Control in Isolated AC Microgrids

Concurrent Cyber Deception Attack Detection of Consensus Control in Isolated AC Microgrids

Observer-Based Security Fuzzy Control for Nonlinear Networked Systems Under Weighted Try-Once-Discard Protocol

This paper is concerned with the observer-based security fuzzy control problem for a class of nonlinear networked systems with imperfectly matched membership functions (MFs) under weighted try-once-discard (WTOD) protocol. The nonlinear networked system is constructed based on interval type-2 (IT2) Takagi-Sugeno (T-S) model, which can properly handle parameter uncertainties via lower and upper MFs. To mitigate the communication load, WTOD protocol is exploited to schedule the transmission order of the signals, in which those signals may be subjected to randomly occurring deception attacks. The observer is constructed under unmeasurable premise variables, and then the controller with imperfect matching MFs is designed according to the estimated states. Moreover, the observer-based controller is derived with Lyapunov stability theory such that the stochastic stability of the augmented closed-loop system with predefined disturbance attenuation performance can be guaranteed. Subsequently, sufficient conditions are provided for the desired observer and controller gain matrices based on linear matrix inequalities (LMIs) technique. Numerical simulation example demonstrates the validity of the proposed observer-based security fuzzy control design scheme.

Static output feedback quantized dissipative security control of singular hybrid systems subject to multiple cyber attacks: A dynamic-memory event-triggered strategy

This paper researches the problem of dynamic-memory event-triggered quantized dissipative security control for singular Markovian jump systems (SMJSs) subject to multiple cyber attacks, which comprise denial of service (DoS) attacks and multichannel deception attacks. By employing a set of historically released packets, a dynamic-memory event-triggered strategy is developed subject to DoS attacks which can decrease the occupancy rate of network communication bandwidth and achieve desired dissipative control performance. Furthermore, the closed-loop SMJSs are established by a switched system model on account of the simultaneous appearance of DoS attacks and deception attacks By constructing an improved mode-dependent piecewise Lyapunov–Krasovskii functional and applying singular value decomposition technique, the criteria of exponential admissibility and strict dissipativity for the controlled SMJSs are acquired by right of linear matrix inequalities. The idiographic form of the static output feedback quantized controller is expressed via a valid decoupling method. Finally, the validity of the proposed approach is testified by a numerical example and direct current motor driven load.

Double-layer event-triggered policy of passivity-based output regulation under deception attacks for switched systems

The problem of secure output regulation (SOR) is solved from the passivity viewpoint for the networked switched system (NSS) permitting severely unstable dynamics in the presence of deception attacks (DAs). Dual Double-layer event-triggered policies (DLETPs) are developed to govern the appropriate triggering in a layered manner in conjunction with parameters of passivity, DAs, and switching characteristics. A tolerant switching strategy (TSS) is proposed to permit destabilizing switching outnumbers stabilizing switching and capture the passivity parameter. In the joint design of DLETPs and TSS in SOR for NSSs, the direct quantificational connections are showcased between the system’s stability, candidate Lyapunov function, event-triggering subject to DAs, and switching behaviors through the system’s passivity degree. Simulation results of speed regulation for an aero-engine model provide the effectiveness of the presented control methodology.

Sensor deception attacks against security in supervisory control systems

This paper investigates the problem of synthesizing sensor deception attacks against security in the context of supervisory control of discrete-event systems (DES). We consider a DES plant controlled by a supervisor, whose security requirement is to maintain the initial-secret of the system undetected, i.e., it does not want to reveal the fact that it was initiated from a secret state. On the other hand, there exists an active attacker that can tamper with the observations received by the supervisor by, e.g., hacking into the communication channel between the sensors and the supervisor. The objective of the attacker is to deceive the supervisor such that the initial-secret is revealed due to incorrect control actions. We investigate the problem from the attacker’s point of view and focus on synthesizing attack strategies that threaten the security of the system. We consider two levels of success of the attacker: one requires that the attacker can detect the initial-secret of the system “almost surely” and the other only requires that the attacker has the possibility to detect the initial-secret of the system. For both cases, we present algorithms for synthesizing successful attack strategies. Our approach is based on the All Attack Structure (AAS) which records state estimates for both the supervisor and the attacker. Structural properties of the security requirements are also leveraged to reduce the synthesis complexity. A running academic example is provided to illustrate the proposed synthesis procedures.

Security-guaranteed filter design for discrete-time Markovian jump delayed systems subject to deception attacks and sensor saturation

This work is devoted the problem of a security-guaranteed filter design for a class of discrete-time Markov jump systems that are vulnerable to stochastic deception attacks and have random sensor saturation. Deception attacks, in particular, are taken into account in the filter when the attacker attempts to modify the broadcast signal in communication networks by inserting some misleading information data into the assessment output. The Bernoulli distribution is satisfied by two sets of introduced stochastic variables. It shows the likelihood that the broadcaster’s data transmissions will be the focus of deception attacks and sensor saturation. The Lyapunov functional technique is established, and criteria are derived to ensure that the system is mean-square stable. Furthermore, explicit expression of the filter gains is obtained by solving a set of linear matrix inequalities. Lastly, two simulation examples including a synthetic genetic regulatory network are provided to further demonstrate the validity and efficiency of the suggested theoretical results.