This paper focuses on the security challenges in cyber-physical systems, emphasizing actuator false data injection attacks. Current post-detection frameworks detect anomalies only after compromised control signals are input into the system and cannot be employed directly at the actuator’s end due to the unavailable innovation. To address this, we introduce a control signal pre-detection framework by a multi-step-ahead prediction method, where a remote estimator engages in multi-step state predictions after calculating the current-time state estimate. Subsequently, the controller formulates a multi-step control strategy and transmits it to the actuator for execution. In light of this architecture, we have devised a detector on the actuator’s side to identify potential threats originating from actuator attacks. Furthermore, we undertake an in-depth analysis to characterize the consequential impacts of various attack patterns on the estimation performance, considering the presence of the proposed detector. Simulations are provided to represent and validate our findings visually.

This paper investigates intelligent reflecting surfaces (IRS)-assisted secure transmission for remote state estimation in the presence of an eavesdropper. To counter the eavesdropper and enhance data reception for the remote estimator, the sensor can dynamically switch between traditional and IRS-assisted channels, the latter combining both traditional and IRS-based links. Compared to traditional channels, IRS-assisted channels reduce eavesdropping rates and improve reception. However, frequent use of IRS-assisted channels increases detection risk by the eavesdropper due to enhanced channel awareness. Thus, a trade-off exists between transmission performance and stealthiness. We propose a framework for channel selection and a metric to evaluate the stealthiness of IRS-based channels. Additionally, the trade-off is modeled as a constrained Markov decision process (MDP), and we derive a sufficient condition for the optimal channel selection policy. By transforming the constrained MDP into an unconstrained one using the Lagrange multiplier approach, we elucidate the optimal policy structure for the unconstrained MDP with discounted costs. Finally, a longitudinal flight example is presented to validate the theoretical results.

Event-based state estimation for linear Gaussian systems has garnered significant attention in recent years, with deterministic and stochastic event-based protocols being the most representative. Previous studies have demonstrated that deterministic event-based protocols outperform stochastic ones in the trade-off between communication rate and estimation performance. However, our research reveals that the performance loss in deterministic event-based protocols can surpass that of stochastic event-based protocols for remote state estimation under specific attack scenarios. We explore the impact of insider attacks and derive a closed-form expression for the estimation error covariance under both protocols. Then, we propose a method for designing the attack threshold to meet stealthiness constraints. For scalar cases, we prove that under the same communication rates, the estimation performance of the stochastic event-based estimator outperforms that of its deterministic counterpart under insider attacks. Numerical simulations corroborate that the empirical results align with the theoretical findings.

Connected autonomous vehicles (CAVs) platoons rely on V2V communication and onboard sensing to maintain safe inter-vehicle spacing, yet cyberattacks on links and sensors can inject false data and destabilize platoon control. This paper proposes a hierarchical Secure Tiny Model Predictive Control (Secure TinyMPC) framework for real-time resilient platooning. A verifiable secret sharing (VSS)-based security layer distributes state shares across communication links to reconstruct trusted states and to support distributed, online detection and estimation of communication and radar attacks. A low-computation TinyMPC control layer then uses these trusted states to rapidly compute control inputs suitable for resource-limited onboard hardware. Simulation studies under diverse communication attacks and radar measurement tampering demonstrate accurate spacing regulation, fast speed convergence, and effective real-time distributed attack detection.

Several research have been adopted to designrobustorstochasticmodel predictive control (MPC) strategies to tackle the model uncertainty. However, the worst-case events do not occur frequently, and the statistical information is neglected, which may render robust MPC strategies conservative. In addition, the stochastic MPC controller design, which computes the probability of constraint satisfaction or performance improvement using expected indicators, cannot rule out the incident of less probability to reduce the possibly catastrophic consequences. For instance, severe variations in wind speed can affect the airflow into the engine, impacting the compression and combustion processes, which may cause a surge or choke phenomenon. In this paper, we take risk-aware strategies into account for MPC design to tackle the tails of a probability distribution that can contribute to moderating the risks. Utilizing the proposed risk measurement method, we establish a risk-constrained MPC framework to seek tradeoffs between improving the control performance of interest and restricting the risk of catastrophic incidents, and we reformulate the proposed risk-constrained MPC in a favorable form that is computationally tractable. Moreover, to ensure the safety requirement of the turbofan engine operating processes, we explicitly analyze the closed-loop properties, e.g., recursive feasibility and stability, of the proposed risk-constrained MPC method. Lastly, we demonstrate the efficiency of the proposed risk-constrained MPC algorithm on the safety-critical control of turbofan engines using real-world data to verify the availability of the proposed algorithm.Note to Practitioners—Model predictive control (MPC) is fundamental for several applications in the process industry, transportation, and robotics. However, due to the existence of the model uncertainty, there may occur severe constraint violations or even instability for certain safety-critical systems, which regard normally given safety constraints as a priority in the deployment of control algorithm. For instance, in the safety-critical control of turbofan engine, the latent force of the dynamical system can be affected by some external surrounding factors, such as inlet wind speed, ambient temperature, barometric pressure, inlet conditions, etc., which may cause surge or choke phenomenon. To tackle the model uncertainty, several pieces of research have been adopted to design MPC strategies in the presence of uncertainties. Nonetheless, the worst-case events do not occur frequently, and the statistical information is neglected, which rendersrobustMPC strategies extremely conservative. In addition, thestochasticMPC controller design, which computes the probability of constraint satisfaction or performance improvement using expected indicators, does not rule out the incident of less probability to reduce the possibly catastrophic consequences. In view of the threat of the safety-critical control of turbofan engines, a risk measurement method and a risk-constrained model predictive control strategy are proposed. Meanwhile, to guarantee the safety and reliability of turbofan engines, we explicitly analyze the proposed algorithm’s closed-loop properties.

This paper addresses the problem of global stabilization in nonlinear time-delay systems with unknown measurement sensitivity. Notably, our system allows for the unknown measurement sensitivity to be non-differentiable, coupled with unmeasurable states, which necessitates the development of observers and control input strategies based on dynamic gain. To optimize resource usage and mitigate network congestion, we introduce an event-triggering mechanism based on two events. This mechanism evaluates dynamic gain and control signals, ensuring a guaranteed positive lower bound on execution time. Moreover, the dynamic gain is designed to compensate for the impact of execution errors. The introduction of a relational sensitivity error allows the unknown measurement sensitivity to converge to a small range. By selecting appropriate Lyapunov-Krasovskii functionals, we eliminate the influence of time-delay, ultimately proving global stability of the closed-loop system. Consequently, comparative simulation results validate the effectiveness of the proposed scheme. Note to Practitioners—This paper explores event-triggered control of nonlinear systems with applications in intelligent transportation, robotics, and aerospace. Notably, we address three critical issues. Firstly, we propose methods to manage unmeasurable states, which is essential for systems like autonomous vehicles and drones where full state measurement is often not feasible. Secondly, we improve existing event-triggering mechanisms to optimize network resource usage, significantly reducing communication frequency, which is particularly beneficial in networked control systems, such as smart grids and industrial automation. Thirdly, we tackle the challenge of unknown measurement sensitivity, relevant for systems operating in uncertain environments like industrial robots and aerospace applications where sensor accuracy can vary. Comparative simulations show that our dynamic event-triggered control method effectively reduces network burden, proving its practical value in real-world applications where network bandwidth and reliability are crucial. In future research, we will also consider reducing the transmission frequency from sensor to controller to further improve system performance. Additionally, we aim to explore adaptive mechanisms to better manage uncertainties in measurement sensitivity, thereby expanding the practical applicability of our approach.

This article discusses the security concerns related to networked systems, where the sensor sends the local estimate to the remote estimator, which may be attacked. Traditionally, in the remote state estimation with the innovation or raw measurement case, denial of service (DoS) and false data injection (FDI) attacks are investigated thoroughly. Notably, for remote state estimation with local estimate cases considered in this article, most existing works consider DoS attacks but not FDI attacks, negatively affecting remote state estimation performance. Furthermore, current detection mechanisms encounter challenges when identifying such attacks due to the unavailable innovation or raw measurement. As such, we study FDI attacks under this framework and provide the corresponding secure analysis using a dual-perspective approach. Specifically, we propose a detector to detect such attacks using the prior information extracted from the remote estimator. Then, we analyze the existence of stealthy attacks and characterize the corresponding performance evaluation for the remote estimation under such attacks. Following this, we construct the optimal attack scheme, maximizing the expected average and terminal estimation error covariances, respectively. To reduce the above vulnerability, we develop a co-design transmission strategy and offer an analytical detection performance evaluation under different attack scenarios. Finally, simulations are provided to illustrate the proposed results.

This paper considers the distributed nonconvex optimization problem, where the goal is to minimize the average of local nonconvex cost functions through local information exchange. Firstly, we propose a distributed optimization algorithm that integrates the gradient tracking method with a dynamic event-triggered communication scheme, thereby reducing communication overhead. Secondly, we demonstrate that the algorithm linearly converges to the global optimum under the Polyak-Łojasiewicz condition, which indicates that every stationary point is a global minimizer. The numerical experiment is presented to validate the theoretical results and confirm the algorithm's effectiveness.

In this article, the problem of sensor transmission power control for remote state estimation is considered, where packet drops may occur. In most existing literature, the task is to choose the transmission power level for sending the latest data packet at each time step. However, only transmitting the latest packet may not fully compensate the estimation performance degradation due to the drops of historical packets. Therefore, in this paper, we propose a new sensor transmission power control scheme, where the power for sending the latest packet in the fixed sensor transmission power control scheme is split into two parts: one for the transmission of the latest packet and the other one for the re-transmission of the historical packet. Specifically, we derive an explicit recursion expression of the estimation error covariance and the steady-state analysis of the proposed scheme. Additionally, we provide a sufficient condition such that the proposed scheme outperforms the fixed sensor transmission power control scheme. Moreover, the analytical properties of the proposed scheme and further results are derived from the special case of sensor scheduling problems. Finally, numerical simulations illustrate the validity of the proposed theoretical methods.Note to Practitioners—There are some scenarios in practice where the packet dropped in the transmission, resulting in degraded estimation and control performance. However, only sending current data in most literature may not compensate for the estimated performance degradation caused by historically dropped data. Motivated by this, we consider a new sensor transmission power control scheme using both drops of historical packets and the current information, which further improves the accuracy of the state estimation. In addition, we allocate the transmission power wisely to send the current data and the drops of historical packets for efficient use of energy due to the limited transmission power in practice. Specifically, we allocate part of the transmission power of the current data to the drops of historical packets for transmission when the packet is dropped under the same energy constraint. The effectiveness of the scheme is illustrated based on simulation study.

In this article, we develop two invariance principles for nonlinear discrete-time switched systems based on multiple Lyapunov functions and multiple weak Lyapunov functions, respectively, which allow the first differences of multiple weak Lyapunov functions to be positive on certain sets. It is shown that the solution of the system is attracted to the largest weakly invariant set in a certain specific region. Then, based on the invariance principle developed and geometrical dissipativity, we obtain the generalized output synchronization for discrete-time dynamical networks with nonidentical nodes by an appropriate switching among several communication topologies. Finally, two examples are provided to demonstrate the effectiveness of the main results.

This article considers malicious attacks on a stochastic event-based state estimation where a smart sensor equipped with the standard Kalman filter is utilized to transmit the local estimate. A novel attack strategy called stealthy insider attack is proposed, which can compromise remote state estimation by hacking the scheduler, reversing the triggering condition, and tampering with the schedule parameter. The discovery of the complete Gaussian crater (CGC) distribution is significant for analyzing various properties of the innovation under the stochastic event-triggered scheme (ETS). An extended CGC distribution is developed to explore the probability distribution of innovation sequences with successive packet losses, and a closed-form expression is derived for the estimation error covariance under attack. Furthermore, to bypass the communication rate detector, a method is presented for tampering with the schedule parameter based on the ergodicity of the underlying Markov chain. Finally, two numerical simulations demonstrate the efficacy of the proposed attack strategy in diminishing the estimation performance of the remote estimator.

This article explores the security control in a remote control system driven by local and remote controllers. By utilizing the information of the local controller (namely, local information, including its mean and error variance), we propose a new actuator-side detector that can prevent performance degradation caused by attacks on the remote control signal, which is transmitted to the actuator through wireless communication. Besides, it can also overcome the difficulties when a standard Kullback–Leibler divergence detector fails to detect such attacks before the control signal is input into the system due to the unavailability of innovation$z_{k}$or measurement$y_{k}$. Subsequently, we characterize the corresponding impacts of different attack patterns on the estimation performance under the proposed detector. Based on this, we offer a compensation mechanism to improve the performance of the remote estimator under the attack. Finally, simulations are provided to illustrate the developed results.

In this article, a dynamic event-triggered adaptive antidisturbance (ETAAD) switching control strategy is proposed for switched systems subject to multisource disturbances. The disturbances are divided into two categories: the available unmodeled disturbance and the unavailable dynamic neural network modeled disturbance. First, a dynamic ET criterion is set based on the system state. Then, a novel dynamic ETA disturbance estimator is introduced to observe the modeled disturbance. Furthermore, according to the ET rule and adaptive disturbance observer, a switched controller is designed. Next, under the controller and switching criterion with the average dwell time limitation, sufficient conditions are given to force the switched systems to realize multisource disturbance suppression (DS), trajectory tracking, and communication resource (CR) saving simultaneously. Meanwhile, the Zeno phenomenon may be caused by the ET rule being excluded. In addition, the presented ETAAD approach is also applicable to the nonswitched systems case. Finally, a simulation case is given to validate the effectiveness of the dynamic ETAAD switching control method.

rgb0.00,0.00,0 Turbofan engine performance optimization is usually formulated as a single objective, closed-form optimization problem by employing a prior mechanism model with an additive, user-preference weight. However, in practical scenarios, the conventional single objective performance optimization may not satisfy the high-performance requirements. For instance, pursuing high-effective thrust will lead to high-turbine inlet temperature due to generating extra heat. Moreover, the system model may be inaccurate or even unavailable, mainly due to the degradation factor, manufacturing tolerance, or time-intensive experiments. Traditionally, the multiobjective optimization methods may require a certain amount of function evaluations, or the convergence properties may not be guaranteed explicitly. To tackle the above-mentioned issues, we formulate the performance optimization of turbofan engines as a multiobjective optimization problem and construct a gradient-free framework to deal with the issue of an inaccurate/unavailable turbofan engine model. Then, to ensure the safety requirement of the turbofan engine operating processes, a multiobjective optimization algorithm is proposed utilizing a gradient-free method, termed Hessian aware gradient estimation-based randomized search (HAGE-RS), and we analyze the corresponding convergence properties of the solved candidates. Finally, we illustrate the proposed algorithm on benchmarks and the performance optimization problem using real-world turbofan engine data under different operating conditions to show superior performance.

Cyber–physical systems (CPSs) driven by a local controller and a remote controller have been gaining significant research interest in recent years due to its application scenarios in practice, such as unmanned aerial vehicles (UAVs). In this article, we consider the security issue in the remote control system with a local controller under stealthy attacks. Under this framework, one controller is designed locally based on the limited measurements collected by a local sensor, and the other controller is designed remotely and is transmitted to the actuator through a wireless communication channel, which may suffer malicious attacks due to its openness. To defend attacks on remote control signal, the K–L divergence-based detector or$\chi ^{2}$detector is often adopted. However, there may be attackers adopting stealthy attacks, which can bypass such detectors. Therefore, we analyze the existence of such attacks, and analytically characterize the worst-estimation performance degradation induced by the remote control signal attack. Further, we construct the optimal attack signal to achieve the upper bound on the estimation performance degradation. In addition, we also give countermeasures against such stealthy attacks. Simulations are provided to illustrate the proposed results.

This article analyzes a security problem for industry cloud control systems (CCSs), where a malicious false data injection (FDI) attacker may deteriorate the system performance by tampering with the data being transmitted. To improve the robustness of CCSs against FDI attacks, a redundancy-based sensor configuration scheme is provided through analysis of the observability under attacks of the nodes in CCSs. Then, a defending resource allocation scheme is developed based on a two-stage Stackelberg game in order to optimize the overall defense capability of the CCSs with limited defending resource. In this two-stage Stackelberg game, the defender of the CCSs acts first and allocates the defending resource to secure the measurements of wireless sensors. With the knowledge of the defender’s strategy, the attacker then decides which target nodes to launch attacks on. The optimal resource allocation strategy is then designed in the sense of the Stackelberg equilibrium. Furthermore, the defense problem under different attacking resource constraints is investigated. Numerical examples illustrate the effectiveness of the proposed approach.

This article aims to investigate the data-driven attack detection and identification problem for cyber-physical systems under sparse actuator attacks, by developing tools from subspace identification and compressive sensing theories. First, two sparse actuator attack models (additive and multiplicative) are formulated and the definitions of I/O sequence and data models are presented. Then, the attack detector is designed by identifying the stable kernel representation of cyber-physical systems, followed by the security analysis of data-driven attack detection. Moreover, two sparse recovery-based attack identification policies are proposed, with respect to sparse additive and multiplicative actuator attack models. These attack identification policies are realized by the convex optimization methods. Furthermore, the identifiability conditions of the presented identification algorithms are analyzed to evaluate the vulnerability of cyber-physical systems. Finally, the proposed methods are verified by the simulations on a flight vehicle system.

Aeroengine performance optimization rem- ains significant for both efficiency and safety during specific operating conditions. Previous works usually solve this optimization problem under a single-objective optimization framework, while multiple objectives need to be optimized simultaneously. Besides, the underlying optimization process requires a variety of function evaluations, and the evaluation cost for an aeroengine is expensive. In reality, the aeroengine model has multiple information sources with different costs and accuracy. The different costs and accuracy of the multiple information sources should be traded off to guide the search for the optimal in a cost-efficient way. Therefore, we propose a multi-information-source framework for enabling efficient multiobjective Bayesian optimization. We construct the surrogate model with a multifidelity Gaussian process and choose the location–source pair with a modified acquisition function. Finally, we apply the proposed method to improve the performance indexes of the aeroengine, which confirms the efficiency of the proposed algorithm.

We study the denial-of-service (DoS) attack power allocation optimization in a multiprocess cyber–physical system (CPS), where sensors observe different dynamic processes and send the local estimated states to a remote estimator through wireless channels, while a DoS attacker allocates its attack power on different channels as interference to reduce the wireless transmission rates, and thus degrading the estimation accuracy of the remote estimator. We consider two attack optimization problems. One is to maximize the average estimation error of different processes, and the other is to maximize the minimal one. We formulate these problems as Markov decision processes (MDPs). Unlike the majority of existing works where the attacker is assumed to have complete knowledge of the CPS, we consider an attacker with no prior knowledge of the wireless channel model and the sensor information. To address this uncertainty issue and the curse of dimensionality, we provide a learning-based attack power allocation algorithm stemming from the double deep Q-network (DDQN) method. First, with a defined partial order, the maximal elements of the action space are determined. By investigating the characteristic of the MDP, we prove that the optimal attack allocations of both problems belong to the set of these elements. This property reduces the entire action space to a smaller subset and speeds up the learning algorithm. In addition, to further improve the data efficiency and learning performance, we propose two enhanced attack power allocation algorithms which add two auxiliary tasks of MDP transition estimation inspired by model-based reinforcement learning, i.e., the next state prediction and the current action estimation. Experimental results demonstrate the versatility and efficiency of the proposed algorithms in different system settings compared with other algorithms, such as the conventional value iteration, double Q-learning, and deep Q-network.

In this article, two algorithms are proposed to solve multiobjective path-constrained dynamic optimization problems. In each algorithm, an adaptive$\varepsilon $-constraint method is employed to solve the multiobjective dynamic optimization problems (MODOPs) with path constraints in two iterative loops. In the outer loop, the adaptive$\varepsilon $-constraint method adaptively adjusts the choice of the parameters$\varepsilon $, which transfers MODOP into a sequence of single-objective dynamic optimization problems (SODOPs) with extra inequality constraints. In the inner loop, two different algorithms are used to solve the single-objective optimization problems. The first algorithm guarantees that the path constraints can be satisfied with any finite prescribed tolerance by replacing path constraints with a finite number of point constraints. Furthermore, the second algorithm guarantees that the path constraints are rigorously satisfied by enforcing the path constraints at a limited number of time points and by restricting the right-hand side of the path constraints. The proposed algorithms are proven to converge within finite iterations. The effectiveness of the algorithms is verified via numerical studies, along with a comparison to a state-of-the-art algorithm.

This article is concerned with the attack detection issue for a class of nonlinear cyber–physical systems (CPSs) with unknown-but-bounded (UBB) noises. The nonlinear system is linearized by utilizing first-order Taylor expansion with Lagrangian remainder, and an observer based on zonotopic sets is proposed to estimate the system states. Then, a novel encoding–decoding strategy (EDS) is provided to improve the attack detection rate by selecting appropriate parameters. In order to alleviate the impact introduced by malicious attacks, a countermeasure is taken into account when an attack is detected by the abnormal detector. Finally, the hardware-in-the-loop (HIL) simulation is provided to illustrate the effectiveness of the proposed results.

We focus on the problem of asymptotically mean-square stabilization in discrete-time stochastic systems that exhibit plant uncertainty, multiple input delays, and multiplicative noises. Our innovative contributions are described as follows. First, we employ a reduction method to transform the original model into a delay-free auxiliary system, and establish an equivalent proposition for stabilization based on this reformulation. On the basis of the reformulated model, we propose two stabilization criteria for the uncertainty-free case, including both Lyapunov-type and Riccati-type criteria. More generally, we extend the stabilization result to the uncertain model, and propose a necessary and sufficient stabilization criterion utilizing matrix homogeneous polynomials. Finally, we explore the existence and uniqueness of a delay margin under certain structural restrictions, and provide a closed-form representation of this margin.

This article studies a security issue in remote distributed consensus estimation where sensors transmit their measurements to remote estimators via a wireless communication network. The relative entropy is utilized as a stealthiness metric to detect whether the data transmitted through the wireless network are attacked. The performance degradation induced by an attacker that attempts to be stealthy or undetected is analyzed, and the corresponding false-data attack strategy is characterized, which can achieve the maximal integrated mean-square error (IMSE). Finally, the tradeoff between the performance degradation and attack stealthiness level is evaluated through an example.

This article studies the attack detection problem of cyber–physical systems (CPSs) with the event-triggered (ET) mechanism. A switching-based moving-target defense (MTD) strategy is developed to detect malicious false-data-injection (FDI) attacks, which is characterized by the average dwell-time (ADT) property. Even if attacks may remain stealthy when the MTD mechanism is adopted, the impact of these stealthy attacks on the system performance is relatively small when the switching signal is unknown to the adversary. In order to reduce the cost of data transmission, an ET mechanism is added into the system. Furthermore, we prove that there exists a lower bound of the execution interval, which indicates that the Zeno phenomenon is excluded. Finally, numerical examples are employed to illustrate the effectiveness of the main results.

This article proposes a novel attack detection approach based on zonotopes for linear parameter-varying (LPV) systems with unknown-but-bounded (UBB) noises. The following three types of attacks are considered: 1) denial-of-service (DoS) attacks; 2) replay attacks (RAs); and 3) false-data-injection (FDI) attacks. In order to reduce the conservativeness, a free-weighting matrix is introduced, which can be computed by solving an optimization problem. Moreover, the radius of the intersection zonotope can be guaranteed to be limited as well. Furthermore, it is not necessary to acquire the knowledge about the specific type of attack in advance. Finally, a numerical example is given to illustrate the validity of the given method.

In this article, a data-driven design scheme of undetectable false data-injection attacks against cyber-physical systems is proposed first, with the aid of the subspace identification technique. Then, the impacts of undetectable false data-injection attacks are evaluated by solving a constrained optimization problem, with the constraints of undetectability and energy limitation considered. Moreover, the detection of designed data-driven false data-injection attacks is investigated via the coding theory. Finally, the simulations on the model of a flight vehicle are illustrated to verify the effectiveness of the proposed methods.