Self-Healing in Cyber–Physical Systems Using Machine Learning: A Critical Analysis of Theories and Tools (original) (raw)
Related papers
Challenges of Machine Learning Applied to Safety-Critical Cyber-Physical Systems
Machine Learning and Knowledge Extraction, 2020
Machine Learning (ML) is increasingly applied for the control of safety-critical Cyber-Physical Systems (CPS) in application areas that cannot easily be mastered with traditional control approaches, such as autonomous driving. As a consequence, the safety of machine learning became a focus area for research in recent years. Despite very considerable advances in selected areas related to machine learning safety, shortcomings were identified on holistic approaches that take an end-to-end view on the risks associated to the engineering of ML-based control systems and their certification. Applying a classic technique of safety engineering, our paper provides a comprehensive and methodological analysis of the safety hazards that could be introduced along the ML lifecycle, and could compromise the safe operation of ML-based CPS. Identified hazards are illustrated and explained using a real-world application scenario—an autonomous shop-floor transportation vehicle. The comprehensive analys...
Artificial intelligence in cyber physical systems
AI & SOCIETY, 2020
This article conducts a literature review of current and future challenges in the use of artificial intelligence (AI) in cyber physical systems. The literature review is focused on identifying a conceptual framework for increasing resilience with AI through automation supporting both, a technical and human level. The methodology applied resembled a literature review and taxonomic analysis of complex internet of things (IoT) interconnected and coupled cyber physical systems. There is an increased attention on propositions on models, infrastructures and frameworks of IoT in both academic and technical papers. These reports and publications frequently represent a juxtaposition of other related systems and technologies (e.g. Industrial Internet of Things, Cyber Physical Systems, Industry 4.0 etc.). We review academic and industry papers published between 2010 and 2020. The results determine a new hierarchical cascading conceptual framework for analysing the evolution of AI decision-maki...
Enhancing Critical Infrastructure Resilience through Machine Learning: A Comprehensive Overview
Enhancing Critical Infrastructure Resilience through Machine Learning, 2024
Critical infrastructure systems, including those in energy, transportation, water management, and telecommunications, are vital components of modern society, providing essential services that support economic prosperity, public safety, and quality of life. However, these systems face a myriad of challenges, ranging from natural disasters and cyber threats to aging infrastructure and population growth. In this context, machine learning (ML) has emerged as a powerful tool for enhancing the resilience of critical infrastructure systems. By analyzing vast amounts of data, detecting patterns, and making predictive insights, ML can help infrastructure operators and stakeholders anticipate, mitigate, and respond to disruptions more effectively. One of the primary advantages of ML is its ability to analyze diverse sources of data to identify vulnerabilities, predict potential failures, and optimize system performance. Despite its potential benefits, the widespread adoption of ML in infrastructure resilience efforts faces several challenges and considerations. These include concerns related to data privacy, algorithm robustness, interpretability, and the need for domain expertise. Ensuring data privacy is crucial, especially when dealing with sensitive information such as infrastructure vulnerabilities or operational data. Algorithm robustness is essential to withstand adversarial attacks, data drift, or model degradation over time. Interpretability is necessary to foster trust and understanding among stakeholders, particularly in safety-critical domains. Moreover, domain expertise is vital to ensure that ML solutions are tailored to the specific needs and challenges of critical infrastructure systems. Addressing these challenges requires interdisciplinary collaboration, stakeholder engagement, and continuous innovation. Interdisciplinary collaboration brings together researchers, practitioners, and policymakers from diverse fields, including computer science, engineering, urban planning, policy, and social sciences, to develop holistic solutions that address the complex challenges of infrastructure resilience. Stakeholder engagement ensures that ML solutions are aligned with the needs and priorities of infrastructure operators, end-users, and communities. Continuous innovation drives progress in ML research and development, leading to the development of more advanced algorithms, techniques, and applications for infrastructure resilience. In conclusion, machine learning holds immense potential for enhancing the resilience of critical infrastructure systems in an interconnected and rapidly evolving world. By leveraging ML algorithms, autonomous systems, and robotics, infrastructure operators and stakeholders can analyze data more effectively, predict and mitigate disruptions, and optimize system performance. However, realizing the full benefits of ML requires addressing challenges related to data privacy, algorithm robustness, interpretability, and domain expertise. This necessitates interdisciplinary collaboration, stakeholder engagement, and continuous innovation to develop responsible, effective, and sustainable ML solutions for infrastructure resilience. As we navigate the complexities of the modern world, the importance of leveraging ML to strengthen critical infrastructure resilience cannot be overstated.
ArXiv, 2020
Modern algorithms in the domain of Deep Reinforcement Learning (DRL) demonstrated remarkable successes; most widely known are those in game-based scenarios, from ATARI video games to Go and the StarCraft~\textsc{II} real-time strategy game. However, applications in the domain of modern Cyber-Physical Systems (CPS) that take advantage a vast variety of DRL algorithms are few. We assume that the benefits would be considerable: Modern CPS have become increasingly complex and evolved beyond traditional methods of modelling and analysis. At the same time, these CPS are confronted with an increasing amount of stochastic inputs, from volatile energy sources in power grids to broad user participation stemming from markets. Approaches of system modelling that use techniques from the domain of Artificial Intelligence (AI) do not focus on analysis and operation. In this paper, we describe the concept of Adversarial Resilience Learning (ARL) that formulates a new approach to complex environment...
Self-Healing Systems: Application and Methodologies-A Review
International Journal of Research, 2020
Self-healing in software applications is patterned after the human cells which regenerates after a damage has been done to it. There are always attacks on software applications that sometimes render the user helpless, since most users are not technicians. If these applications will be able to recover from attacks and get back to normal state before it was attaced without letting the user know that such attack ever happened, a self-healing mechanism has been achieved in that application. In this paper, we tried to look at what is self-healing, methodologies that some researchers have proposed in order to achieve self-healing in any given system, system faults and its remedies, self-healing life cycle, and applications of self-healing in a system.
Self-healing systems: Foundations and challenges
Self-Healing and Self-Adaptive Systems, Germany. Dagstuhl Seminar Proceedings, 2009
Abstract. The term and characteristic of self-healing, applied to systems, is often seen from different fields of computer science, such as fault tolerance or network and service management, with diverging semantics. Since this impression was confirmed also during the first discussions of the Dagstuhl seminar on” Self-Healing and Self-Adaptive Systems”, a seminar's working group on” Terminology” was formed with the objective to address the question of finding commonalities and differences in a self-healing characteristic of stand- ...
Cyber-Physical Systems Resilience: State of the Art, Research Issues and Future Trends
ArXiv, 2019
Ideally, full integration is needed between the Internet and Cyber-Physical Systems (CPSs). These systems should fulfil time-sensitive functions with variable levels of integration with their environment, incorporating data storage, computation, communications, sensing, and control. There are, however, significant problems emerging from the convergence between CPS and Internet of Things (IoT) areas. The high heterogeneity, complexity, and dynamics of these resource-constrained systems bring new challenges to their robust and reliable operation, which implies the need for novel resilience management strategies. This paper surveys the state of the art in the relevant fields and, discusses the research issues and future trends that emerge. Thus, we hope to provide new insights into the management of resilient CPSs, formed by IoT devices, modelled by Game Theory, and flexibly programmed using the latest software and virtualization platforms.