Sensor Attacks Research Articles

Secure state estimation for cyber-physical systems by unknown input observer with adaptive switching mechanism

A new state estimation method is proposed in this manuscript for a class of linear cyber-physical systems (CPSs) with sparse sensor attacks, unknown input and output disturbances. The sensor attack will make part of the measurement signal received by the remote observer inaccurate. In order to achieve the secure state estimation (SSE) for the investigated linear CPSs, an unknown input observer (UIO) with adaptive switching mechanism is designed. Inspired by the existing results, a set of sub-observers is designed which can exclude the influence of unknown input and output disturbances. To achieve switching between different sub-observers, the adaptive switching mechanism is designed according to the switching adversarial principle. Some parameters of the observer are obtained by designing a set of linear matrix inequalities (LMIs). The sufficient condition for the existence of the observer and the proof of its effectiveness are respectively given by theoretical analysis. Finally, the effectiveness of the proposed method is proved by two Matlab simulation experiments.

Event-Triggered Tracking Control of Nonlinear Systems Under Sparse Attacks and its Application to Rigid Aircraft

This study is aimed at addressing the event-based state estimation and tracking control problems for a class of nonlinear systems, the output of which is measured by multiple sensors, but the adversary can manipulate nearly half of the measurements simultaneously. First, a sampled-data-based event-triggered strategy is developed to reduce unnecessary data transmissions under sparse sensor attacks, and the transmitted data are filtered by a data selector to obtain reliable data. Subsequently, an output-prediction-based continuous-discrete observer is improved so that it can estimate continuous-time system states from the event-triggered output, rather than being limited to time-triggered sampled output. Further, to design a tracking controller with the segmentally differentiable estimated states, a backstepping method incorporating tracking differentiators is proposed. Finally, the effectiveness of the proposed method is demonstrated by applying it in the simulation of a rigid aircraft.

Finite time secure state estimation for linear cyber‐physical systems with actuator and sensor attacks

AbstractThis manuscript studies the problem of secure state estimation (SSE) for a class of linear discrete cyber‐physical systems (CPSs) considered with output disturbances, actuator, and sensor attacks. The disturbances and attacks considered are unbounded. A finite time SSE (FTSSE) method is designed to realize SSE for CPSs in a predetermined finite time. First, a finite time observer is designed to implement the FTSSE for CPSs with actuator attacks and output disturbances. After that, the FTSSE method for CPSs with simultaneous sensor attacks is proposed by extending the designed finite time observer. The necessary conditions for the existence of the proposed FTSSE method are given, and its effectiveness is analyzed theoretically. In addition, the simultaneous estimation of disturbance and attack can be achieved by the designed estimator when the system satisfies some given conditions. Finally, the effectiveness of the designed method is verified by a simulation experiment.

Event-Triggered Adaptive Security Path Following Control for Unmanned Ground Vehicles Under Sensor Attacks

This paper is concerned with the event-triggered adaptive security control (ET-ASC) for the path following of networked unmanned ground vehicles (UGVs) subject to sensor attacks. Firstly, a security model is well established to capture the dynamics of path following control of UGVs under sensor attacks. Then, the ET-ASC with respect to correction signal is proposed to mitigate the effects of such sensor attacks. In what follows, two theorems, which include both input to state stability (ISS) criterion and controller design method, are carefully derived for the path following control of UGVs under the proposed ET-ASC scheme. The advantage of the proposed ET-ASC scheme lies in that it can actively amend the sensor attacks in an adaptive way and exclude Zeno phenomenon naturally. At last, a verification simulation experiment are conducted to show the effectiveness of the proposed ET-ASC method.

Event-triggered secure control of nonlinear multi-agent systems under sensor attacks

In this paper, a security consistent tracking control scheme with event-triggered strategy and sensor attacks is developed for a class of nonlinear multi-agent systems. For the sensor attacks on the system, a security measurement preselector and a state observer are introduced to combat the impact of the attacks and achieve secure state estimation. In addition, command filtering technology is introduced to overcome the “complexity explosion” caused by the use of the backstepping approach. Subsequently, a new dynamic event-triggered strategy is proposed, in which the triggering conditions are no longer constants but can be adjusted in real time according to the adaptive variables, so that the designed event-triggered mechanism has stronger online update ability. The measurement states are only transmitted through the network based on event-triggered conditions. The proposed adaptive backstepping algorithm not only ensures the security of the system under sensor attacks but also saves network resources and ensures the consistent tracking performance of multi-agent systems. The boundedness of all closed-loop signals is proved by Lyapunov stability analysis. Simulation examples show the effectiveness of the control scheme.

Adaptive self-triggered control-based cooperative output regulation of heterogeneous multi-agent systems under sensor and actuator attack

The cooperative output regulation problem for heterogeneous multi-agent systems under sensor and actuator attacks with an adaptive self-triggering strategy is considered in this article. The attack considered is the false data injection attack. To resolve the system, a dynamic internal model is introduced for each agent, and a distributed adaptive self-triggering control scheme is proposed. The results show that intermittent communication is possible for all agents using the proposed control mechanism and no agent shows Zeno behavior. The adaptive controller is used to guarantee that the regulated output for the heterogeneous multi-agent systemss is achieved with cooperative uniform ultimate boundedness under both actuator and sensor attacks. Given the whole state of information is unavailable, the adaptive dynamic system is designed based on receivable compromised or uncompromised output information. The effectiveness of the given control scheme is illustrated by two examples.

An adaptive control scheme for switched nonlinear cyber‐physical systems against state‐dependent sensor attacks and input‐dependent actuator attacks

AbstractThis article considers the problem of adaptive control against deception attacks for a class of switched nonlinear cyber‐physical systems (CPSs), in which each subsystem has more general and unknown nonlinearities. Specifically, an adaptive controller is designed for CPSs with unknown switching mechanisms to mitigate the impact of state‐dependent sensor attacks and input‐dependent actuator attacks. Compared with the existing researches, the actuator attacks considered in our article are input‐dependent, which means the controller is substantially attacked, besides, the signs of unknown time‐varying gains caused by state‐dependent sensor attacks and input‐dependent actuator attacks are all unknown. To deal with these scenarios, Nussbaum‐type functions are introduced. In addition, by constructing a common Lyapunov function for all subsystems, the closed‐loop system signals are proved to be globally bounded under arbitrary switchings. Finally, we give a simulation example of a continuously stirred tank reactor system with state‐dependent sensor attacks and input‐dependent actuator attacks to illustrate the effectiveness of our results.

A Circuit-Level Solution for Secure Temperature Sensor

Temperature sensors play an important role in modern monitoring and control applications. When more and more sensors are integrated into internet-connected systems, the integrity and security of sensors become a concern and cannot be ignored anymore. As sensors are typically low-end devices, there is no built-in defense mechanism in sensors. It is common that system-level defense provides protection against security threats on sensors. Unfortunately, high-level countermeasures do not differentiate the root of cause and treat all anomalies with system-level recovery processes, resulting in high-cost overhead on delay and power consumption. In this work, we propose a secure architecture for temperature sensors with a transducer and a signal conditioning unit. The proposed architecture estimates the sensor data with statistical analysis and generates a residual signal for anomaly detection at the signal conditioning unit. Moreover, complementary current–temperature characteristics are exploited to generate a constant current reference for attack detection at the transducer level. Anomaly detection at the signal conditioning unit and attack detection at the transducer unit make the temperature sensor attack resilient to intentional and unintentional attacks. Simulation results show that our sensor is capable of detecting an under-powering attack and analog Trojan from a significant signal vibration in the constant current reference. Furthermore, the anomaly detection unit detects anomalies at the signal conditioning level from the generated residual signal. The proposed detection system is resilient against any intentional and unintentional attacks, with a detection rate of 97.73%.

Resilient set-based state estimation for linear time-invariant systems using zonotopes

This paper considers the problem of set-based state estimation for linear time-invariant (LTI) systems under time-varying sensor attacks. Provided that the LTI system is stable and observable via every single sensor and that at least one sensor is uncompromised, we guarantee that the true state is always contained in the estimated set. We use zonotopes to represent these sets for computational efficiency. However, we show that intelligently designed stealthy attacks may cause exponential growth in the algorithm’s worst-case complexity. We present several strategies to handle this complexity issue and illustrate our resilient zonotope-based state estimation algorithm on a rotating target system.

A virtual models-based CAVs platoon resilient to network and sensor attacks

The most typical topological structure of Connected and Autonomous cars (CAVs) is a platoon formation, in which vehicles move closely together, following the leader’s instructions, and offer autonomous and intelligent transport services. Nevertheless, several attacks that inject false data into networks and onboard sensors may endanger platoon stability, resulting in severe collisions over road networks in the worst case. This paper proposes a mechanism for cooperative CAV platoon recovery in the face of false data injection attacks called PReCAV. PReCAV employs virtual and physical system models to maintain the resilience of the CAV platoon despite network and sensor attacks. An evaluation by NS3 and SUMO simulations showed that PReCAV provided stability and resilience to the CAV platoon formation in a representative scenario face to false data injection attacks in the network and the sensors. For injecting false data into the network and sensors, we defined an attack model that manipulates information, such as speed and localization, using a uniform distribution to vary the falsification size over time, that is, how far does the information value vary from the correct one. During network and sensors attacks, PReCAV showed average values of 0.16 m/s2 acceleration, 21.76 m/s of speed, and 28.35 m of the distance between vehicles in the platoon when the desired speed was 22 m/s. Meanwhile, for a desired speed of 15 m/s, PReCAV maintained average values of 0.11 m/s2, 15.26 m/s, and 22.42 m in the platoon. Those results demonstrate the ability of PReCAV to maintain vehicle dynamics in the platoon in the face of attacks.

A Novel Standalone Implementation of MDNN Controller for DC–DC Converter Resilient to Sensor Attacks—A Design Approach

Power electronic converters have become an integral part of the direct current microgrid (DCMG) systems. The efficient control of these converters will decide the performance of the DC microgrids. With the evolution of cyber physical systems, all these power converters are integrated into the communication networks to achieve intelligent and smart control. Data in the communication networks are highly vulnerable towards cyber-attacks, leaving it unaddressed will lead to substantial economic losses and disasters. This article proposes a standalone implementation of multi-deep neural network (MDNN) based DC-DC converter and its application to detect false data injection (FDI) attacks at the sensor level. A MDNN is developed with a combination of deep neural network (DNN) and error detection network (EDN). DNN is used as the controller to achieve closed loop operation of DC-DC converter and EDN is used to detect and mitigate the FDI attack. Initially, proposed scheme is executed in MATLAB simulink platform with various disturbances at controller level and several attack scenarios are verified with its experimental results.

Distributed secure consensus control of nonlinear multi-agent systems under sensor and actuator attacks

In this paper, we focus on an output secure consensus control issue for nonlinear multi-agent systems (MASs) under sensor and actuator attacks. Followers in an MAS are in strict-feedback form with unknown control directions and unknown dead-zone input, where both sensors and nonlinear characteristics of dead-zone in actuators are paralyzed by malicious attacks. To deal with sensor attacks, uncertain dynamics in individual follower are separated by a separation theorem, and estimation parameters are introduced for compensating and mitigating the influence from adversaries. The influence from actuator attacks are treated as a total displacement in a dead-zone nonlinearity, and an upper bound, as well as its estimation, is introduced for this displacement. The dead-zone nonlinearity, sensor attacks and unknown control gains are gathered together regarded as composite unknown control directions, and Nussbaum functions are utilized to address the issue of unknown control directions. A distributed secure consensus control strategy is thus developed recursively for each follower in the framework of surface control method. Theoretically, the stability of the closed-loop MAS is analyzed, and it is proved that the MAS achieves output consensus in spite of nonlinear dynamics and malicious attacks. Finally, theoretical results are verified via a numerical example and a group of electromechanical systems.

Defending State-Feedback Based Controllers Against Sensor Attacks

Context: This paper is motivated by the need to improve the resilience of industrial control systems. Many control systems currently operating in the industry were designed and implemented before the boom in communications (wired and wireless networks) within industrial control systems. However, nowadays they operate connected to the communications network. This increase in connectivity has made the systems susceptible to cyber-attacks that seek to degrade the proper operation of the control loop even affecting only one sensor.
 Method: We use concepts from fault tolerant control and classic control theory to show that it is possible to reconstruct the system state without (any) one of the system outputs. This fact is used in the recalculation of the control signal through an algorithm of attack detection and isolation, to avoid for an attack to be feedback to the system, mitigating its effect. We show the effectiveness of our proposal with simulations on a four-tanks testbed using Matlab and Simulink.
 Results: We show that a bank of unknown input observers can be designed to recover true information from attacked sensors, i.e., recover the information without the effect of the attack. Therefore, the estimation obtained from those observers can be utilized for computing a control action that mitigates the effect of the attack.
 Conclusions: This mitigation of the attack prevents a single sensor attack from signi cantly impair the action of low-level controllers, improving the resilience on the system just modifying the digital controller architecture. The development presented is limited to cyber-attacks on system sensors pre- sented one at a time, that still can compromise seriously the system behavior. Future work will address the extension of the presented results in situations with simultaneous attacks on more than one sensor, and/or consider attacks on the control system actuators.

DSC‐based finite‐time adaptive resilient control for a class of nonstrict‐feedback switched nonlinear systems with deception attacks

AbstractIn this paper, the finite‐time adaptive resilient control problem for a class of switched nonlinear systems with deception attacks by using the dynamic surface control (DSC) strategy is addressed. A more general class of uncertain nonstrict‐feedback switched nonlinear systems with the sensor and actuator deception attacks are considered. Deception attacks can damage sensor, control and switching data, which results in the switching signal and the finite‐time control gains unknown, meanwhile the conventional dynamic surface errors inacessible. To overcome the difficulty, first, the approximation technique of neural networks (NNs) is used to deal with the unknown nonlinear terms and compensate the sensor and actuator deception attacks. Secondly, a new coordinate transformation and a novel switched adaptive finite‐time filter are designed to mitigate the effects caused by the sensor attacks and avoid the problem of “explosion of complexity”. Thirdly, Nussbaum technology is exploited to handle the unknown finite‐time control gains problem caused by deception attacks. Then, by the proposed control strategy, the practical finite‐time stability analysis is presented. Finally, a simulation example is provided to validate the feasibility of the proposed control method.

On decidability of existence of nonblocking supervisors resilient to smart sensor attacks

Cybersecurity of discrete event systems (DES) has been gaining more and more attention recently, due to its high relevance to the so-called 4th industrial revolution that heavily relies on data communication among networked systems. One key challenge is how to ensure system resilience to sensor and/or actuator attacks, which may tamper data integrity and service availability. In this paper we focus on some key decidability issues related to smart sensor attacks. We first present a sufficient and necessary condition that ensures the existence of a smart sensor attack, which reveals a novel demand–supply relationship between an attacker and a controlled plant, represented as a set of risky pairs. Each risky pair consists of a damage string desired by the attacker and an observable sequence feasible in the supervisor such that the latter induces a sequence of control patterns, which allows the damage string to happen. It turns out that each risky pair can induce a smart weak sensor attack. Next, we show that, when the plant, supervisor and damage language are regular, it is possible to remove all such risky pairs from the plant behaviour, via a genuine encoding scheme, upon which we are able to establish our key result that the existence of a nonblocking supervisor resilient to smart sensor attacks is decidable. To the best of our knowledge, this is the first result of its kind in the DES literature on cyber attacks. The proposed decision process renders a specific synthesis procedure that guarantees to compute a resilient supervisor whenever it exists, which so far has not been achieved in the literature.

Robust predictability in discrete event systems under sensor attacks

The problem of robust predictability against sensor attacks is investigated. The objective of a diagnoser is to predict the occurrence of a critical event of a discrete event system (DES) under partial observation. An attacker may rewrite the diagnoser observation by inserting fake events or erasing real events. Two novel structures, namely, real diagnoser and the fake diagnoser, are constructed based on the diagnoser of the system. We compute the hybrid diagnoser as the parallel composition of the real diagnoser and the fake diagnoser. The hybrid diagnoser can be used to verify if a critical event of the system is robustly predictable when an attacker tampers with the diagnoser observation.

An adaptive control scheme against state‐dependent sensor attacks and input‐dependent actuator attacks in cyber‐physical systems

AbstractThe application of cyber‐physical systems (CPSs) in many areas is threatened by its embedded structure. This paper considers the design of an adaptive controller for linear CPSs against deception attacks. Specifically, a corrective signal is designed to suppress or offset the impact of sensor and actuator attacks, especially when the controller and the correction signal are also attacked. First, an adaptive controller is developed for time‐invariant deception attacks and further extended to time‐varying deception attacks. In each case, state‐independent sensor attacks and input‐independent actuator attacks, as well as state‐dependent sensor attacks and input‐dependent actuator attacks, are discussed. For the considered closed‐loop system with time‐invariant state‐dependent sensor attacks and input‐dependent actuator attacks, the controller the authors developed ensures the asymptotic stability of the systems. For the other three cases of the corresponding system, the controllers are also designed, which ensures uniform ultimate boundedness. Finally, several numerical examples are given to explain the availability of our control scheme.

Resilient Event-Triggered Model Predictive Control for Adaptive Cruise Control Under Sensor Attacks

Dear Editor, This letter addresses the resilient model predictive control (MPC) problems for adaptive cruise control (ACC) systems under sensor attacks. In the light of vulnerabilities of ACC systems to sensor attacks, an intrusion detection mechanism is proposed at the controller side to distinguish abnormal data. Then, the robust control gains are derived to design the terminal region constraint for MPC. Further, an improved event-triggered scheme is presented, which shows the ability to relieve the computation burden in the implementation of MPC, by combining the memory-based event-triggered condition and the maximum allowed time interval condition. Finally, efficacy validations are given in simulations.

GAN-Based Robust Motion Planning for Mobile Robots Against Localization Attacks

Motion planning (MP) is essential but challenging for mobile robots. Most of the existing MP methods, at each instant, compute an action based on the states of the robot and the surrounding obstacles, assuming that the robot's localization module is attack-free. Unfortunately, the localization module is vulnerable to sensor attacks, such as GPS spoofing attacks. In this letter, we propose a novel robust framework, <monospace xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">GAN-MP</monospace> , where a generative adversarial network (GAN) is exploited to mitigate the localization attacks, and the state-of-the-art MP methods are applied to generate collision-free actions. Specifically, <monospace xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">GAN-MP</monospace> aims to learn a Generator to compute the potential positions of the robot. Consequently, it can reserve the robot's benign states while correcting the attacked states. Hence, it is suitable for benign and attacked scenarios without any attack detector. In addition, <monospace xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">GAN-MP</monospace> is method-agnostic and can be easily integrated with any existing MP method. We instantiate <monospace xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">GAN-MP</monospace> with a deep reinforcement learning method to demonstrate its design and training processes. Comprehensive experiments show that <monospace xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">GAN-MP</monospace> can mitigate localization attacks and guarantee safe motion. We also demonstrate the robustness and generalization of <monospace xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">GAN-MP</monospace> .

Synthesizing Attack-Aware Control and Active Sensing Strategies Under Reactive Sensor Attacks

We consider the probabilistic planning problem for a defender (P1) who can jointly query the sensors and take control actions to reach a set of goal states while being aware of possible sensor attacks by an adversary (P2) who has perfect observations. To synthesize a provably-correct, attack-aware joint control and active sensing strategy for P1, we construct a stochastic game on graph with augmented states that include the actual game state (known only to the attacker), the belief of the defender about the game state (constructed by the attacker based on his knowledge of defender's observations). We present an algorithm to compute a belief-based, randomized strategy for P1 to ensure satisfying the reachability objective with probability one, under the worst-case sensor attack carried out by an informed P2. We prove the correctness of the algorithm and illustrate using an example.