A Digital Twin and Machine Learning-Based Framework for Assessing and Enhancing the Cyber Resilience of Critical Infrastructures in Smart Cities

A Digital Twin and Machine Learning-Based Framework for Assessing and Enhancing the Cyber Resilience of Critical Infrastructures in Smart Cities

MahboubehMoradtajari1

1) Department of Computer Engineering, National Skill University, Tehran, Iran. Email:

IOI Code : AAHD-ABGKAK
Publication : National Conference on Information Technology, Nanotechnology, Artificial Intelligence and Technological Futures Studies(itnaf.ir)
Abstract :
    The rapid expansion of smart cities and the increasing dependence of critical infrastructures on digital technologies, the Internet of Things (IoT), industrial control systems, and communication networks have made cyber threats one of the most significant challenges facing urban management. Cyberattacks can disrupt the operation of critical infrastructures such as power grids, intelligent transportation systems, water and wastewater facilities, healthcare centers, and telecommunication networks, resulting in significant economic, social, and security consequences. Under these circumstances, enhancing cyber resilience and improving the ability to predict, detect, and manage cyber threats have become essential requirements for the sustainable development of smart cities.This study proposes an integrated framework based on Digital Twin technology and Machine Learning for assessing and enhancing the cyber resilience of critical infrastructures in smart cities. In the proposed framework, a digital twin of the target infrastructure is first developed to recreate operational conditions, events, and interactions among system components within a virtual environment. Subsequently, various cyberattack scenarios are simulated to generate the required operational and security data, which are then analyzed using the Random Forest algorithm. This process enables the identification of cyber threats, anomalies, and potential vulnerabilities within the infrastructure.To evaluate the performance of the proposed model, accuracy, confusion matrix analysis, and feature importance analysis are employed. The results indicate that the integration of Digital Twin technology and Machine Learning provides an effective approach for threat detection, attack scenario analysis, and cyber resilience assessment. Furthermore, the proposed framework can serve as a valuable decision-support tool for urban managers, infrastructure planners, and security operation centers in the development of resilient smart cities.

Keywords : digital twin machine learning cyber resilience smart cities critical infrastructure cybersecurity random forest