Critical Cyber-physical Systems Research Articles

A Probabilistic Approach to Adaptive Protection in the Smart Grid

Smart grids are critical cyber-physical systems that are vital to our energy future. Smart grids’ fault resilience is dependent on the use of advanced protection systems that can reliably adapt to changing grid conditions. The vast amount of operational data generated and collected in smart grids can be used to develop these protection systems. However, given the safety-criticality of protection, the algorithms used to analyze this data must be stable, transparent, and easily interpretable to ensure the reliability of the protection decisions. Additionally, the protection decisions must be fast, selective, simple, and reliable. To address these challenges, this paper proposes a data-driven protection strategy, based on Gaussian Discriminant Analysis (GDA), for fault detection and isolation. This strategy minimizes the communication requirements for time-inverse relays, facilitates their coordination, and optimizes their settings. The interpretability of the protection decisions is a key focus of this paper. The method is demonstrated by showing how it can protect the medium-voltage CIGRE network as it transitions between islanded and grid-connected modes and radial and mesh topologies.

Using Artificial Intelligence in the Security of Cyber Physical Systems

The prominence of cyber security continues to increase on a daily basis. Following the cyber attacks in recent years, governments have implemented a range of regulations. The advancement of technology and digitalization has led to the creation of new vulnerabilities that cyber attackers can exploit. The digitalization of facilities such as energy distribution networks and water infrastructures has enhanced their efficiency, thereby benefiting states and society. The modern sensors, controllers, and networks of these new generation facilities have made them susceptible to cyber attackers. While all forms of cyber attacks are detrimental, targeting critical cyber-physical systems presents a heightened level of peril. These assaults have the potential to disrupt the social structure and pose a threat to human lives. Various techniques are employed to guarantee the security of these facilities, which is of utmost importance. This study examined the applications of machine learning and deep learning methods, which are sub-branches of artificial intelligence that have recently undergone a period of significant advancement. Intrusion detection systems are being created for the networks that facilitate communication among the hardware components of the cyber-physical system. Another potential application area involves the development of models capable of detecting anomalies and attacks in the data generated by sensors and controllers. Cyber physical systems exhibit a wide range of diversity. Due to the wide range of variations, it is necessary to utilize specific datasets for training the model. Generating a dataset through attacks on a functional cyber-physical system is unattainable. The study also analyzed the solutions to this problem. Based on the analyzed studies, it has been observed that the utilization of artificial intelligence enhances the security of cyber physical systems.

Threat modeling in smart firefighting systems: Aligning MITRE ATT&CK matrix and NIST security controls

Industrial automation technologies are envisioned as multi-device systems that are constantly interacting with one another and with enterprise systems. In these industrial systems, the industrial internet of things (IIoT) significantly improves system efficiency, scalability, ease of control, and monitoring. These benefits have been achieved at the cost of greater security risks, thus making the system vulnerable to cyberattacks. Historically, industrial networks and systems lacked security features like authentication and encryption due to intended isolation over the Internet. Lately, remote access to these IIoT systems has made an attempt of holistic security alarmingly critical. In this research paper, a threat modeling framework for smart cyber–physical system (CPS) is proposed to get insight of the potential security risks. To carry out this research, the smart firefighting use case based on the MITRE ATT&CK matrix was investigated. The matrix analysis provided structure for attacks detection and mitigation, while system requirement collection (SRC) was applied to gather generic assets’ information related to hardware, software and network. With the help of SRC and MITRE ATT&CK, a threat list for the smart firefighting system was generated. Conclusively, the generated threat list was mapped on the national institute of standards and technology (NIST) security and privacy controls. The results show that these mapped controls can be well-utilized for protection and mitigation of threats in smart firefighting system. In future, critical cyber–physical systems can be modeled upon use case specific threats and can be secured by utilizing the presented framework.

Gaps in Asset Management Systems to Integrate Railway Companies’ Resilience

Railways and metros are safe, efficient, reliable, and environmentally friendly mass carriers. They are critical cyber-physical systems (CPS) that are attractive targets for cyber and/or physical attacks. SAFETY4RAILS project delivers methods and systems to increase the safety and resilience of track-based inter-city railway and intra-city metro transportation. Asset management plays a fundamental role in resilience management. This study analyses the gaps in asset management systems of rail infrastructure. The objective of the study is to understand the weaknesses and vulnerabilities in an asset management system that impacts resilience. The form of triangulation fashion was used for the analysis of consequences for each threat event. The research conducted included: a systematic literature review; a multiple case study review; and an analysis. The strength of asset inventory, condition inspection methods and decision-making scenarios were analysed, and as an expanded part of this analysis, mitigation actions linked to the vulnerabilities were identified. The study implies that asset management systems are most important in resilience management’s response and recovery phases where the largest sudden economic implications can take place. The results of the gap analysis could be used to provide policy recommendations and standardisation efforts.

Thinking in Systems, Sifting Through Simulations: A Way Ahead for Cyber Resilience Assessment

The interaction between the physical world and information technologies creates advantages and novel emerging threats. Cyber-physical systems (CPSs) result vulnerable to cyber-related disruptive scenarios, and, for some critical systems, cyber failures may have fallouts on society and environment. Traditional risk analysis in no more sufficient to deal with these problems. New techniques are gaining increasing consensus, especially those based on systems theory. In this context, the System-Theoretic Process Analysis for Security (STPA-Sec) extends the Systems-Theoretic Accident Modelling and Processes (STAMP) model considering cyber threats, and identifying unsafe and unsecure controls throughout a cyber socio-technical system. Despite its large usage as a descriptive tool, there is still limited use of STPA-Sec in (semi-)quantitative terms. This article presents System-Theoretic Process Analysis for Security with Simulations (STPA-Sec/S), a methodological interface between STPA-Sec and quantitative resilience assessment based on simulation models. The methodology is instantiated in a demonstrative case study of a water treatment plant, and its critical CPSs which may impact both community health, and environment. The obtained results show how STPA-Sec/S foster systems understanding, allow a systematic identification of its major criticalities, and the respective quantification.

A conversation with Heena Rathore

Ubiquity's senior editor Dr. Bushra Anjum chats with Dr. Heena Rathore, assistant professor in the Computer Science Department at Texas State University, on how cyber-physical systems (CPS) are inherently similar to humans in their reliance upon sensing their environments and making decisions. They further discuss how this similarity motivates the exploration of novel AI algorithms inspired by the principles of cognition and neuroscience to ensure the security of critical CPS infrastructure, such as IoT networks, medical devices, and connected vehicles.

Adoption of ACPS in Nuclear Reactor Analysis

Nuclear power plants are a paramount example of critical cyber-physical systems. Some of the current researches in nuclear reactor analysis concern the degree of acceptable uncertainty of the whole system. Some difficulties arise from weaving fields of different disciplines, such as computer science (e.g. embedded software and hardware), mechatronics (e.g. sensors and actuators) and physics (e.g. neutronics and thermal-hydraulics). To complicate the scenario further, each field demands different disciplines, competencies and different teams. In this short paper, we highlight the importance of the cross-fertilization of different disciplines in the nuclear reactor domain and we investigate emerging methods to control uncertainty in the nuclear reactor design.

Don’t Drink the Cyber: Extrapolating the Possibilities of Oldsmar’s Water Treatment Cyberattack

Water treatment represents an essential critical infrastructure sector which has a direct impact on the health and well-being of its customers. Water treatment is often performed by municipalities with very limited budgets for cybersecurity resources. These underfunded, high-impact, targets represent an emerging cyber warfare attack-surface paradigm which poses a direct threat to the quality of life for millions of people. On February 5th, 2021, a water treatment plant in Oldsmar, Florida was the victim of an attempted cyberattack. This attack commanded the system to add a dangerous amount ofsodium hydroxide to water which supplied thousands. Direct exposure to sodium hydroxide causes painful burns to the exposed area with permanent internal damage likely upon ingestion. A system operator noticed this malicious behaviour and corrected the situation, minimizing the attack’s impact. This paper outlines the attack and illustrates how minor modifications to the attacker’s tactics, techniques, and procedures could have resulted in a cyber-derived catastrophe for thousands of unsuspecting citizens. Lastly, this paper explores the effectiveness of various low-cost cyber-physical security technologies when pitted against differing attacker models in these theoretical scenarios. These cybersecurity solutions are evaluated by cost, ease of use, implementation difficulty, and ability to support safe operation continuity when faced with adversary behaviour. The results of this evaluation illuminate a path forward for low-cost threat mitigation which increases the difficulty to compromise these critical cyber-physical systems. With attacks targeting industrial control systems on the rise, the Oldsmar water treatment cyberattack represents more than an individual incident, it can be viewed as a reflection of the current status of thousands of similar critical infrastructure systems that have yet to be caught in crosshairs of a competent and willing adversary with financial incentives and cyber warfare mission requirements serving as impetus for adversary willingness and any resulting large-scale cyber cataclysm.

A Survey of Crypto Ransomware Attack Detection Methodologies: An Evolving Outlook.

Recently, ransomware attacks have been among the major threats that target a wide range of Internet and mobile users throughout the world, especially critical cyber physical systems. Due to its unique characteristics, ransomware has attracted the attention of security professionals and researchers toward achieving safer and higher assurance systems that can effectively detect and prevent such attacks. The state-of-the-art crypto ransomware early detection models rely on specific data acquired during the runtime of an attack’s lifecycle. However, the evasive mechanisms that these attacks employ to avoid detection often nullify the solutions that are currently in place. More effort is needed to keep up with an attacks’ momentum to take the current security defenses to the next level. This survey is devoted to exploring and analyzing the state-of-the-art in ransomware attack detection toward facilitating the research community that endeavors to disrupt this very critical and escalating ransomware problem. The focus is on crypto ransomware as the most prevalent, destructive, and challenging variation. The approaches and open issues pertaining to ransomware detection modeling are reviewed to establish recommendations for future research directions and scope.

From LTL to rLTL monitoring: improved monitorability through robust semantics

Runtime monitoring is commonly used to detect the violation of desired properties in safety critical cyber-physical systems by observing its executions. Bauer et al. introduced an influential framework for monitoring Linear Temporal Logic (LTL) properties based on a three-valued semantics for a finite execution: the formula is already satisfied by the given execution, it is already violated, or it is still undetermined, i.e., it can still be satisfied and violated by appropriate extensions of the given execution. However, a wide range of formulas are not monitorable under this approach, meaning that there are executions for which satisfaction and violation will always remain undetermined no matter how it is extended. In particular, Bauer et al. report that 44% of the formulas they consider in their experiments fall into this category. Recently, a robust semantics for LTL was introduced to capture different degrees by which a property can be violated. In this paper we introduce a robust semantics for finite strings and show its potential in monitoring: every formula considered by Bauer et al. is monitorable under our approach. Furthermore, we discuss which properties that come naturally in LTL monitoring—such as the realizability of all truth values—can be transferred to the robust setting. We show that LTL formulas with robust semantics can be monitored by deterministic automata, and provide tight bounds on the size of the constructed automaton. Lastly, we report on a prototype implementation and compare it to the LTL monitor of Bauer et al. on a sample of examples.

Resilience Management of Critical Cyber-Physical Systems: A Multiple Case Study Analysis

Modern societies are highly dependent on different critical cyber-physical systems (CPS). The growth of software (= cyber) layer, in size and percentage of the overall system is a future trend. Our society’s critical CPS — cyber, energy, water, transportation and communication — lacks of resilience, typically losing essential functionality following adverse events. This multiple case study research analyses five prior research projects from the safety and security field. The target of the paper is to research how resilience management of critical cyber-physical systems can be understood.

Meta-Learning to Improve Unsupervised Intrusion Detection in Cyber-Physical Systems

Artificial Intelligence (AI)- based classifiers rely on Machine Learning (ML) algorithms to provide functionalities that system architects are often willing to integrate into critical Cyber-Physical Systems (CPSs) . However, such algorithms may misclassify observations, with potential detrimental effects on the system itself or on the health of people and of the environment. In addition, CPSs may be subject to threats that were not previously known, motivating the need for building Intrusion Detectors (IDs) that can effectively deal with zero-day attacks. Different studies were directed to compare misclassifications of various algorithms to identify the most suitable one for a given system. Unfortunately, even the most suitable algorithm may still show an unsatisfactory number of misclassifications when system requirements are strict. A possible solution may rely on the adoption of meta-learners, which build ensembles of base-learners to reduce misclassifications and that are widely used for supervised learning. Meta-learners have the potential to reduce misclassifications with respect to non-meta learners: however, misleading base-learners may let the meta-learner leaning towards misclassifications and therefore their behavior needs to be carefully assessed through empirical evaluation. To such extent, in this paper we investigate, expand, empirically evaluate, and discuss meta-learning approaches that rely on ensembles of unsupervised algorithms to detect (zero-day) intrusions in CPSs. Our experimental comparison is conducted by means of public datasets belonging to network intrusion detection and biometric authentication systems, which are common IDSs for CPSs. Overall, we selected 21 datasets, 15 unsupervised algorithms and 9 different meta-learning approaches. Results allow discussing the applicability and suitability of meta-learning for unsupervised anomaly detection, comparing metric scores achieved by base algorithms and meta-learners. Analyses and discussion end up showing how the adoption of meta-learners significantly reduces misclassifications when detecting (zero-day) intrusions in CPSs.

Analysing Mission-critical Cyber-physical Systems with AND/OR Graphs and MaxSAT

Cyber-Physical Systems (CPS) often involve complex networks of interconnected software and hardware components that are logically combined to achieve a common goal or mission; for example, keeping a plane in the air or providing energy to a city. Failures in these components may jeopardise the mission of the system. Therefore, identifying the minimal set of critical CPS components that is most likely to fail, and prevent the global system from accomplishing its mission, becomes essential to ensure reliability. In this article, we present a novel approach to identifying the Most Likely Mission-critical Component Set (MLMCS) using AND/OR dependency graphs enriched with independent failure probabilities. We address the MLMCS problem as a Maximum Satisfiability (MaxSAT) problem. We translate probabilities into a negative logarithmic space to linearise the problem within MaxSAT. The experimental results conducted with our open source tool LDA4CPS indicate that the approach is both effective and efficient. We also present a case study on complex aircraft systems that shows the feasibility of our approach and its applicability to mission-critical cyber-physical systems. Finally, we present two MLMCS-based security applications focused on system hardening and forensic investigations.

Complex System Governance as a Foundation for Enhancing the Cybersecurity of Cyber-Physical Systems

This article investigates the possible benefits of complex system governance (CSG) as a foundation for enhancing cybersecurity in critical cyber-physical systems (CPS). CPS are intrinsically linked to cyberspace and vulnerable to a wide range of risks stemming from physical and cyber threats. There remains a lack of robust frameworks for addressing the issue of cybersecurity for CPS at the metasystem level. In response, the authors suggest CSG as an organizing construct capable of providing a greater degree of cohesion and as a means to provide for design, execution, and evolution of ‘metasystemic' functions necessary to provide for communication, control, and coordination, and integration is critical to the cybersecurity of CPS. In this article, CSG is introduced as a potential construct for enhancing cybersecurity in CPS. A hypothetical case study application is then provided to illustrate the potential for this research. Finally, the authors offer conclusions and suggest future research. Keywords Case Study, Complex System Governance, Cyber-Physical Systems, Management Cybernetics, Metasystem Function, Power Grid, Systems Thinking

Protection of large-scale smart grids against false data injection cyberattacks leading to blackouts

Power system is one of the most critical cyber–physical systems in a sustainable society, whose security and reliability can significantly impact other infrastructures. Therefore, security and resilience of smart grids are one of the most challenging research questions in protection of critical infrastructures. After cyberattacks on the U.S. power grid in 2018, the national security agency announced cyberattacks might cause a blackout in transmission systems. Therefore, it is critical to model cyberattack scenarios, propose a detection strategy to mitigate these threats in a realistic context, and study how the models can improve the defense of large-scale grids. In this paper, a bi-level mixed-integer linear programming (BMILP) model is developed to accurately model false data injections (FDIs) that are targeted to overflow multiple transmission lines and cause a blackout in large-scale grids. Compared to the existing research, the proposed model considers that attackers might have limited access to measurement buses and models attacks on targeted transmission lines without being detected by existing DC state-estimation. In addition, to protect the system against these attacks, it is proved that a detection framework based on recursive weighted least-square (WLS) state-estimation can detect the FDIs, unlike classical weighted least square (WLS) estimation that fails to identify stealthy FDIs. To validate the effectiveness of the proposed attack model and detection framework in practical grid infrastructures, the IEEE 118-bus benchmark and a 2000-bus synthetic grid replicating the electricity network of Texas, U.S. are used.

Assessing IoT enabled cyber-physical attack paths against critical systems

Internet of Things (IoT) increase the interconnectivity and interoperability of systems in various critical sectors, such as industrial control, healthcare and smart transportation systems. At the same time, as IoT technologies enable systems to interact both in cyber and physical ways, they also act as enablers of complex attack paths against critical systems. In this paper we propose a novel risk-based methodology for identifying and assessing IoT-enabled attack paths against critical cyber-physical systems. While the majority of existing approaches focus on cyber system connectivity only, the proposed methodology models both cyber and physical interactions. In comparison to existing cyber physical approaches that grow exponentially, our approach is significantly more efficient, by utilizing an attack tree topology; the critical system is set as the root (target) of an attack tree that is recursively build, based on the identified cyber-physical system interactions. Our methodology uses well-known building blocks such Common Vulnerabilities and Exposures (CVE), Common Vulnerability Scoring System (CVSS) and threat modeling. Furthermore,we significantly reduce false positives by prioritizing the identified attack paths in a risk manner, which, in turn, can assist decision makers in effectively mitigating multi-hop attack paths. To validate our methodology, we developed a proof-of-concept implementation and tested it using a realistic scenario from the healthcare sector. Our results show that the proposed methodology can efficiently identify and assess hidden and/or underestimated cyber physical attack paths.

A Deep Learning Ensemble for Network Anomaly and Cyber-Attack Detection.

Currently, expert systems and applied machine learning algorithms are widely used to automate network intrusion detection. In critical infrastructure applications of communication technologies, the interaction among various industrial control systems and the Internet environment intrinsic to the IoT technology makes them susceptible to cyber-attacks. Given the existence of the enormous network traffic in critical Cyber-Physical Systems (CPSs), traditional methods of machine learning implemented in network anomaly detection are inefficient. Therefore, recently developed machine learning techniques, with the emphasis on deep learning, are finding their successful implementations in the detection and classification of anomalies at both the network and host levels. This paper presents an ensemble method that leverages deep models such as the Deep Neural Network (DNN) and Long Short-Term Memory (LSTM) and a meta-classifier (i.e., logistic regression) following the principle of stacked generalization. To enhance the capabilities of the proposed approach, the method utilizes a two-step process for the apprehension of network anomalies. In the first stage, data pre-processing, a Deep Sparse AutoEncoder (DSAE) is employed for the feature engineering problem. In the second phase, a stacking ensemble learning approach is utilized for classification. The efficiency of the method disclosed in this work is tested on heterogeneous datasets, including data gathered in the IoT environment, namely IoT-23, LITNET-2020, and NetML-2020. The results of the evaluation of the proposed approach are discussed. Statistical significance is tested and compared to the state-of-the-art approaches in network anomaly detection.

Development of the concept of electronic government construction in the conditions of synergetic threats

The object of research is the process of forming the concept of building electronic government in the context of the synergy and hybridity of modern threats. The development of high technologies, the rapid growth of information and communication technologies (ICT) and computing resources allow us to improve not only the sphere of life and services for society, but also to create both individual elements and full-scale projects of the digital state. One of the most problematic places in the formation of a digital state and/or e-government remains the security of a single portal and/or registry. In the context of the rapid growth of computing resources, cybercriminals are realizing a combination of threats with social engineering methods, which allows to obtain a synergistic effect and hybridity.An assessment of the capabilities of a quantum computer is obtained, which significantly reduces the level of security when using traditional and asymmetric cryptography. This is due to the fact that the proposed methods for breaking symmetric and asymmetric cryptosystems are implemented in the post-quantum period with polynomial complexity. The paper proposes a hierarchical concept of introducing a governance model based on e-government, which allows deepening the principles of democracy and virtually eliminating corruption with the required level of security. For its deployment, it is necessary to ensure at each level the protection of information and communication and mobile technologies that ensure the functionality of electronic government. The paper studies the main threats to critical cyberphysical systems, as the basis of the mechanisms for fulfilling the functions of e-government. Given the construction of full-scale quantum computers, the main goals and objectives of the concept of building electronic government are formed in the work. Thanks to this, it is possible to obtain the basic principles and functionality at each level of the Concept of the formation of electronic government. Compared with similar well-known approaches, this ensures that the necessary level of not only computer and cyber literacy of the population is taken into account, but also that the state’s cyberspace is fully covered. It allows to provide the functionality of electronic government in the context of the integration of modern threats.

Moving Target Defense Approach to Detecting Stuxnet-Like Attacks

Recent cybersecurity incidents such as Stuxnet and Irongate alert us to the threats faced by critical cyber-physical systems. These attacks compromise the control signals to push the system to unsafe regions and meanwhile, inject fake sensor measurements to cover the ongoing attack. Detecting these Stuxnet-like (SL) attacks still remains an open research issue. This paper analyzes the taxonomy, construction, and implication of SL attacks in CPS control loops. We propose to apply the moving target defense (MTD) approach that actively changes the system configuration to detect SL attacks, since these attacks are generally constructed based on the knowledge about the system's configuration. We analyze the basic conditions for MTD to be successful. Finally, as a case study, we apply MTD for the secondary voltage control of power grids and present simulation results based on the IEEE 39-bus test system under realistic settings.

A System-level Behavioral Detection Framework for Compromised CPS Devices

Cyber-Physical Systems (CPS) play a significant role in our critical infrastructure networks from power-distribution to utility networks. The emerging smart-grid concept is a compelling critical CPS infrastructure that relies on two-way communications between smart devices to increase efficiency, enhance reliability, and reduce costs. However, compromised devices in the smart grid poses several security challenges. Consequences of propagating fake data or stealing sensitive smart grid information via compromised devices are costly. Hence, early behavioral detection of compromised devices is critical for protecting the smart grid’s components and data. To address these concerns, in this article, we introduce a novel and configurable system-level framework to identify compromised smart grid devices. The framework combines system and function call tracing techniques with signal processing and statistical analysis to detect compromised devices based on their behavioral characteristics. We measure the efficacy of our framework with a realistic smart grid substation testbed that includes both resource-limited and resource-rich devices. In total, using our framework, we analyze six different types of compromised device scenarios with different resources and attack payloads. To the best of our knowledge, the proposed framework is the first in detecting compromised CPS smart grid devices with system and function-level call tracing techniques. The experimental results reveal an excellent rate for the detection of compromised devices. Specifically, performance metrics include accuracy values between 95% and 99% for the different attack scenarios. Finally, the performance analysis demonstrates that the use of the proposed framework has minimal overhead on the smart grid devices’ computing resources.