As Industrial Cyber-Physical Systems (ICPS) become increasingly interconnected, they are exposed to a growing array of hybrid threats that target both digital and physical assets. A central challenge lies in identifying and implementing effective security countermeasures that mitigate such threats without disrupting critical operations. In this context, Decision Support Systems (DSS) offer a promising avenue for enhancing incident response. This can be done by assisting security operators in selecting optimal mitigation strategies. However, existing DSS solutions often face significant limitations, including poor scalability, high computational demands, and dependence on simulated attack scenarios that may not reflect real-world conditions. To overcome these challenges, we propose a novel hybrid knowledge-based DSS for intrusion response in ICPSs. Our approach integrates expert-driven evaluations with dynamic risk assessment to ensure both scalability and computational efficiency. Risk is continuously assessed using multiple Bayesian Networks. Moreover, expert knowledge is incorporated through an adapted Analytic Hierarchy Process (AHP) framework, capable of handling incomplete information, to evaluate the effectiveness and operational impact of available countermeasures. A cyber-physical detection module further enhances the system by providing detected events along with their associated detection probabilities. The optimal response strategy is selected by solving a bi-objective optimization problem. In particular, the problem amounts to minimizing both residual risk and the potential negative impact on ICPS functionality. The proposed DSS is validated through a proof-of-concept implementation based on a real-world laboratory case study. This demonstrates its practical applicability and robustness in complex industrial environments.

Federated unlearning (FU) algorithms allow clients in federated settings to exercise their "right to be forgotten" by removing the influence of their data from a collaboratively trained model.Existing FU methods maintain data privacy by performing unlearning locally on the client-side and sending targeted updates to the server without exposing forgotten data; yet they often rely on server-side cooperation, revealing the client's intent and identity without enforcement guarantees -compromising autonomy and unlearning privacy.In this work, we propose EFU (Enforced Federated Unlearning), a cryptographically enforced FU framework that enables clients to initiate unlearning while concealing its occurrence from the server.Specifically, EFU leverages functional encryption to bind encrypted updates to specific aggregation functions, ensuring the server can neither perform unauthorized computations nor detect or skip unlearning requests.To further mask behavioral and parameter shifts in the aggregated model, we incorporate auxiliary unlearning losses based on adversarial examples and parameter importance regularization.Extensive experiments show that EFU achieves nearrandom accuracy on forgotten data while maintaining performance comparable to full retraining across datasets and neural architectures -all while concealing unlearning intent from the server.Furthermore, we demonstrate that EFU is agnostic to the underlying unlearning algorithm, enabling secure, function-hiding, and verifiable unlearning for any client-side FU mechanism that issues targeted updates.

Device heterogeneity poses major challenges in Federated Learning (FL), where resource-constrained clients slow down synchronous schemes that wait for all updates before aggregation. Asynchronous FL addresses this by incorporating updates as they arrive, substantially improving efficiency. While its efficiency gains are well recognized, its privacy costs remain largely unexplored—particularly for high-end devices that contribute updates more frequently, increasing their cumulative privacy exposure. This paper presents the first comprehensive analysis of the efficiency-fairness-privacy trade-off in synchronous vs. asynchronous FL under realistic device heterogeneity. We empirically compareFedAvgand staleness-aware FedAsync using a physical testbed of five edge devices spanning diverse hardware tiers, integrating Local Differential Privacy (LDP) and the Moments Accountant to quantify per-client privacy loss. Using Speech Emotion Recognition (SER) as a privacy-critical benchmark, we show that FedAsync achieves up to 10× faster convergence but exacerbates fairness and privacy disparities: high-end devices contribute 6–10× more updates and incur up to 5× higher privacy loss, while low-end devices suffer amplified accuracy degradation due to infrequent, stale, and noise-perturbed updates. These findings motivate the need for adaptive FL protocols that jointly optimize aggregation and privacy mechanisms based on client capacity and participation dynamics, moving beyond static, one-size-fits-all solutions.

Cyber Situation Awareness requires effective methods to interpret complex, dynamic network data, and Granular Computing offers a powerful framework for managing such complexity through abstraction. In this work, we propose a granular computing-based approach for network activity classification that supports Cyber Situation Awareness by combining the Clustering-by-Time method with the principle of justifiable granularity. The system selects the most informative subsets of traffic within time windows, summarizes them into optimized frames, and trains a Random Forest classifier for anomaly detection. Evaluated on the LUFlow dataset, the approach achieves significant data reduction—up to 98%—while maintaining good detection accuracy. This enables scalable and efficient intrusion detection in complex network environments.

Autonomous wheelchair-drone systems represent a promising advancement in assistive mobility, enabling enhanced navigation in complex and dynamic environments. However, floor surface anomalies—such as uneven terrain, obstacles, and hazardous floor conditions—pose significant challenges to safe and efficient operation. This paper presents a novel approach improving situation awareness and self-adaptation by integrating floor surface-anomaly detection in autonomous wheelchair-drone systems. A specific architecture for situation-awareness is proposed, combining machine learning-based anomaly detection with adaptive motion planning, to enhance the system’s resilience and responsiveness. Experimental results in simulated scenarios using Yolo-based architecture on real-world datasets demonstrate improved anomaly detection performances compared to the state-of-the-art, reducing the risk of instability and improving user safety. Experiments show a mAP50 of 0.764 and a F1 of 0.742 using a YoloV11s architecture. The research presented in this paper has been developed within the European project named REXASI-PRO, which aims to develop trustworthy AI solutions to assist individuals with reduced mobility.

This paper presents a simulation for traffic evacuation during railway disruptions to enhance urban resilience. The research focuses on large-scale railway networks and provides flexible simulation settings to accommodate multiple node or line failures. The evacuation optimization model is mathematically formulated using matrix computation and nonlinear programming. The simulation integrates railway lines operated by various companies, along with external geographical features of the network. Furthermore, to address computational complexity in large-scale graph networks, a subgraph partitioning solution is employed for computation acceleration. The model is evaluated using the extensive railway network of Greater Tokyo. Data collection included both railway network structure and real-world GPS footfall data to estimate the number of station-area visitors for simulation input and evaluation purposes. Several evacuation scenarios were simulated for major stations including Tokyo, Shinjuku, Shibuya and so on. The results demonstrate that both evacuation passenger flow (EPF) and average travel time (ATT) during emergencies were successfully optimized, while remaining within the capacity constraints of neighboring stations and the targeted disruption recovery times.

Parkinson’s disease (PD) is a chronic neurological condition causing an assortment of motor and cognitive prodromes. Each individual’s PD symptoms develop differently due to the variability of the ailment. This study aims to introduce the KNN Imputed Spearman’s Rank and Jaccard Convolutional Deep Neural Learning (KISRJCDNL) technique for automating early PD diagnosis depending on speech analysis. This work enhances disease diagnosis performance through preprocessing and early, precise PD detection. Several information collected from the given dataset are initially taken as input. Then, the preprocessing stage converts raw data into a structured format. Afterward, Spearman’s Rank Feature Selective and Jaccard Index–based Convolutional Deep Neural Learning Classifier with four layers, one input layer, one output layer, and two hidden layers, are deployed for diagnosing PD by efficiently performing the data classification. Experimental evaluation uses the Early Biomarkers of the PD dataset by different factors. Findings support the claim that the proposed KISRJCDNL technique enhances accuracy by 14%, reducing feature selection time, error rate, overall time, and space complexity by 16%, 43%, 36%, and 22% compared to the existing deep learning methods.

Federated Learning (FL) enables model training across decentralized devices by communicating solely local model updates to an aggregation server. Although such limited data sharing makes FL more secure than centralized approached, FL remains vulnerable to inference attacks during model update transmissions. Existing secure aggregation approaches rely on differential privacy or cryptographic schemes like Functional Encryption (FE) to safeguard individual client data. However, such strategies can reduce performance or introduce unacceptable computational and communication overheads on clients running on edge devices with limited resources. In this work, we present EncCluster , a novel method that integrates model compression through weight clustering with recent decentralized FE and privacy-enhancing data encoding using probabilistic filters to deliver strong privacy guarantees in FL without affecting model performance or adding unnecessary burdens to clients. We performed a comprehensive evaluation, spanning various datasets and architectures, to demonstrate EncCluster scalability across encryption levels. Our findings reveal that EncCluster significantly reduces communication costs — below even conventional FedAvg — and accelerates encryption by more than four times over all baselines; at the same time, it maintains high model accuracy and enhanced privacy assurances.

Digital twin development of industrial cyber-physical systems requires modeling, simulation, and monitoring to provide accurate digital replicas mimicking system dynamics. To this aim, modern solutions employ data-driven approaches to capture normal system dynamics using historical data, and anomaly detection using run-time data. However, these approaches are typically process-agnostic and have limited explainability, which negatively impacts trustworthiness. Hence, we propose a novel framework for the digital twin development of industrial cyber-physical systems based on process mining, which connects data-driven and process-based analyses. The framework implements an offline phase that systematically identifies the most suitable process model to capture normal dynamics and simulates faulty dynamics for what-if analysis. The subsequent online phase involves run-time monitoring for anomaly detection. We develop a proof-of-concept application of the framework addressing a water distribution case study that allows physical fault injection. Results highlight the importance of adopting high-quality process models to significantly improve anomaly detection and time performance.

In the era of smart transportation, ensuring the reliability, safety, and security of railway systems is paramount. This paper presents an approach to anomaly detection in smart railways, leveraging self-adaptation, process mining, and digital twins. By integrating real-time data analytics with adaptive algorithms, the framework dynamically identifies and responds to anomalies, enhancing operational efficiency and safety. Extensive data from sensors and IoT devices embedded within the railway infrastructure is continuously monitored and analyzed using advanced machine learning algorithms within digital twins. These algorithms adjust their parameters in real-time to accommodate changes in the operational environment, ensuring robustness and accuracy. Process mining techniques extract valuable insights from both historical and real-time data, identifying patterns and deviations that may indicate potential anomalies. This continuous improvement loop not only detects anomalies but also understands their underlying causes, facilitating effective interventions. Digital twins provide a virtual replica of critical assets, updated in real-time, enabling proactive maintenance and predictive analysis. They also serve as a testing ground for new algorithms and strategies, ensuring risk-free evaluation before deployment. The integration of self-adaptation, process mining, and digital twins significantly improves detection accuracy and response times. Applied to railway operational scenarios, this method effectively identifies anomalies, classifying them into minor deviations or major system faults, leading to enhanced operational efficiency, reduced downtime, and improved safety for passengers and personnel. This approach has been developed within several European projects focused on data-driven innovations for smarter railways, utilizing the latest advancements in artificial intelligence and machine learning.

In this paper, we tackle the challenge of generating synthetic log files using generative adversarial networks to support smart-troubleshooting experimentation. Log files are critical for implementing monitoring systems for smart-troubleshooting, as they capture valuable information about the activities and events occurring within the monitored system. Analyzing these logs is crucial for effective smart-troubleshooting, enhancing the overall efficiency, reliability, and security of smart manufacturing processes. However, accessing public log data is difficult due to privacy concerns and the need to protect sensitive information. Moreover, for the purpose of effective troubleshooting, it is essential to have datasets that include fault, error, and failure logs as well as standard logs. In recent years, synthetic log files have emerged as a promising solution to augment limited real-world datasets and facilitate the development and evaluation of anomaly detection techniques. Building on this concept of synthetic data, we have developed a specific log generation technique and dataset tailored for testing smart-troubleshooting techniques in heteroge-neous connected systems environments, such as industrial cyber-physical systems and the internet of things. First, we propose a methodology that generates synthetic log files based on generative adversarial networks. Later, we instantiate this methodology using different Generative Adversarial Network implementations and present a validation and a comprehensive comparative analysis of their performance. Eventually, we provide a robust dataset for anomaly detection and threat analysis in cyberspace security. Based on the results of our comparison, CTGAN has shown superior performance in generating high-quality synthetic log files.

Accurately identifying and profiling users is one of the primary challenges of many modern web applications. This paper presents an approach for user profiling that utilizes Fuzzy User Signatures combined with browser fingerprinting techniques. Our approach analyzes users' web domain visit frequencies and categories to determine their preferences and behaviors. Fuzzy User Signatures provide a condensed representation of user activities, enabling a framework for assessing user similarity. This method can significantly improve web navigation experiences by allowing for personalized content and product recommendations. The approach has been evaluated on a dataset comprising users' web activities combined with browser fingerprints, achieving overall good performances.

Cloud computing (CC) is a network‐based concept where users access data at a specific time and place. The CC comprises servers, storage, databases, networking, software, analytics, and intelligence. Cloud security is the cybersecurity authority dedicated to securing cloud computing systems. It includes keeping data private and safe across online‐based infrastructure, applications, and platforms. Securing these systems involves the efforts of cloud providers and the clients that use them, whether an individual, small‐to‐medium business, or enterprise uses. Security is essential for protecting data and cloud resources from malicious activity. A cloud service provider is utilized to provide secure data storage services. Data integrity is a critical issue in cloud computing. However, using data storage services securely and ensuring data integrity in these cloud servers remain an issue for cloud users. We introduce a unique piecewise regressive Kupyna cryptographic hash blockchain (PRKCHB) technique to secure cloud services with higher data integrity to solve these issues. The proposed PRKCHB method involves user registration, cryptographic hash blockchain, and regression analysis. Initially, the registration process for each cloud user is performed. After registering user particulars, Davies–Meyer Kupyna’s cryptographic hash blockchain generates the hash value of data in each block. When a user requests data from the server, a piecewise regression function is used to validate their identity. Furthermore, the Gaussian kernel function recognizes authorized or unauthorized users for secure cloud information transmission. The regression function results in original data by enhanced integrity in the cloud. An analysis of the proposed PRKCHB technique evaluates different existing methods implemented in Python. The results contain different metrics: data confidentiality rate, data integrity rate, authentication time, storage overhead, and execution time. Compared to conventional techniques, findings corroborate the assertion that the proposed PRKCHB technique improves data confidentiality and integrity by up to 9% and 9% while lowering storage overhead, authentication time, and execution time by 10%, 12%, and 12%.

Sensor ontology is the kernel technique of the Intelligent Sensor System, which provides a structured framework to organize and interpret the knowledge of the Internet of Things (IoT). However, the ontology heterogeneity issue hampers the communication of sensor ontologies. Sensor Ontology Matching (SOM) can find semantically identical entities between two ontologies, which is an effective method to address this issue. However, due to their complicated semantic relationships, it is a challenge to construct an effective Similarity Feature (SF) to distinguish the heterogeneous sensor entities. Although Evolutionary Algorithms (EAs) based matching techniques have shown their effectiveness in the ontology matching field, they suffer from drawbacks such as high computational complexity and expert-dependent solution evaluation. To overcome these drawbacks, this paper proposes a novel Light Genetic Programming (L-GP) to automatically construct SF for SOM. First, a simplified evolutionary mechanism is designed to improve the efficiency of the SOM process. Second, a novel fitness function based on the approximate evaluation metric is introduced to automatically guide the search direction of L-GP. Lastly, a two-stage tournament selection operator is presented to balance the quality and complexity of the solutions, improving the accuracy of the SOM results. The experiment uses ten pairs of real-world SOM tasks to test the performance of L-GP, and the experimental results show that L-GP significantly outperforms state-of-the-art matching techniques.

Federated learning (FL) as a novel paradigm in Artificial Intelligence (AI), ensures enhanced privacy by eliminating data centralization and brings learning directly to the edge of the user's device. Nevertheless, new privacy issues have been raised particularly during training and the exchange of parameters between servers and clients. While several privacy-preserving FL solutions have been developed to mitigate potential breaches in FL architectures, their integration poses its own set of challenges. Incorporating these privacy-preserving mechanisms into FL at the edge computing level can increase both communication and computational overheads, which may, in turn, compromise data utility and learning performance metrics. This paper provides a systematic literature review on essential methods and metrics to support the most appropriate trade-offs between FL privacy and other performance-related application requirements such as accuracy, loss, convergence time, utility, communication, and computation overhead. We aim to provide an extensive overview of recent privacy-preserving mechanisms in FL used across various applications, placing a particular focus on quantitative privacy assessment approaches in FL and the necessity of achieving a balance between privacy and the other requirements of real-world FL applications. This review collects, classifies, and discusses relevant papers in a structured manner, emphasizing challenges, open issues, and promising research directions.

Context: Federated Learning is an approach to distributed machine learning that enables collaborative model training on end devices. FL enhances privacy as devices only share local model parameters instead of raw data with a central server. However, the central server or eavesdroppers could extract sensitive information from these shared parameters. This issue is crucial in applications like speech emotion recognition (SER) that deal with personal voice data. To address this, we propose Optimized Paillier Homomorphic Encryption (OPHE) for SER applications in FL. Paillier homomorphic encryption enables computations on ciphertext, preserving privacy but with high computation and communication overhead. The proposed OPHE method can reduce this overhead by combing Paillier homomorphic encryption with pruning. So, we employ OPHE in one of the use cases of a large research project (DAIS) funded by the European Commission using a public SER dataset.

Context: Speech Emotion Recognition (SER) is a valuable technology that identifies human emotions from spoken language, enabling the development of context-aware and personalized intelligent systems.To protect user privacy, Federated Learning (FL) has been introduced, enabling local training of models on user devices.However, FL raises concerns about the potential exposure of sensitive information from local model parameters, which is especially critical in applications like SER that involve personal voice data.Local Differential Privacy (LDP) has prevented privacy leaks in image and video data.However, it encounters notable accuracy degradation when applied to speech data, especially in the presence of high noise levels.In this paper, we propose an approach called LDP-FL with CSS, which combines LDP with a novel client selection strategy (CSS).By leveraging CSS, we aim to improve the representatives of updates and mitigate the adverse effects of noise on SER accuracy while ensuring client privacy through LDP.Furthermore, we conducted model inversion attacks to evaluate the robustness of LDP-FL in preserving privacy.These attacks involved an adversary attempting to reconstruct individuals' voice samples using the output labels provided by the SER model.The evaluation results reveal that LDP-FL with CSS achieved an accuracy of 65-70%, which is 4% lower than the initial SER model accuracy.Furthermore, LDP-FL demonstrated exceptional resilience against model inversion attacks, outperforming the non-LDP method by a factor of 10.Overall, our analysis emphasizes the importance of achieving a balance between privacy and accuracy in accordance with the requirements of the SER application.

Level Crossings are among the most critical railway assets, concerning both the risk of accidents and their maintainability, due to intersections with promiscuous traffic and difficulties in remotely monitoring their health status. Failures can be originated from several factors, including malfunctions in the bar mechanisms and warning devices, such as light signals and bells. This paper focuses on the intelligent detection of anomalies in warning bells through non-intrusive acoustic monitoring by: (1) introducing a new concept for autonomous monitoring of level crossings; (2) generating and sharing a specific dataset collecting relevant audio signals from publicly available audio recordings; (3) implementing and evaluating a solution combining deep learning and transfer learning for warning bell detection. The results show a high accuracy in detecting anomalies and suggest viability of the approach in real-world applications, especially where network cameras with on-board microphones are installed for multi-purpose level crossing surveillance.

Modern cyber-physical systems based on the Industrial Internet of Things (IIoT) can be highly distributed and heterogeneous, and that increases the risk of failures due to misbehavior of interconnected components, or other interaction anomalies. In this paper, we introduce a conceptual architecture for IIoT anomaly detection based on the paradigms of Digital Twins (DT) and Autonomic Computing (AC), and we test it through a proof-of-concept of industrial relevance. The architecture is derived from the current state-of-the-art in DT research and leverages on the MAPE-K feedback loop of AC in order to monitor, analyze, plan, and execute appropriate reconfiguration or mitigation strategies based on the detected deviation from prescriptive behavior stored as shared knowledge. We demonstrate the approach and discuss results by using a reference operational scenario of adequate complexity and criticality within the European Railway Traffic Management System.

Abstract The emergence of 5G technologies opens up new opportunities for railway communications. One of the foundational aspects of 5G architecture is its control-plane programmability, which can be achieved through Software Defined Networking (SDN). In railway scenarios, this can be used to dynamically reconfigure the network for a more effective and efficient management of communication flows produced by moving trains. The paper presents a framework for integrating modelling and analysis tools into a programmable control plane specifically tailored to railway communications. We introduce the concept of domain-awareness in the network control plane as an SDN-enabled feature that allows achieving application-specific advantages besides those purely expressed in terms of key performance indicators such as the quality of service. We propose a reference architecture in which domain-awareness in the control plane is obtained by considering information gathered by network devices and ad-hoc communication gateways that are able to detect relevant signalling events. In the architecture, the actual behaviour of the SDN controller is governed by applications that are able to react to specific triggers and re-configure network devices accordingly. We also provide a methodological framework based on model-driven engineering and formal methods, including dynamic state machines, for the automatic generation of SDN control plane logic.

Rotating equipment is considered as a key component in several industrial sectors. In fact, the continuous operation of many industrial machines such as sub-sea pumps and gas turbines relies on the correct performance of their rotating equipment. In order to reduce the probability of malfunctions in this equipment, condition monitoring, and fault diagnosis systems are essential. In this work, a novel approach is proposed to perform fault diagnosis in rotating equipment based on permutation entropy, signal processing, and artificial intelligence. To that aim, vibration signals are employed for an indication of bearing performance. In order to facilitate fault diagnosis, fault detection and isolation are performed in two separate steps. As first, once a vibration signal is received, the faulty state of the bearing is determined by permutation entropy. In case a faulty state is detected, the fault type is determined using an approach based on signal processing and artificial intelligence. Wavelet packet transform and envelope analysis of the vibration signals are utilized to extract the frequency components of the fault. The proposed approach allows for the automatic selection of a frequency band that includes the characteristic resonance frequency of the fault, which is subject to change in different operational conditions. The method works by extracting the proper features of the signals that are used to decide about the faulty bearing’s condition by a multi-output adaptive neuro-fuzzy inference system classifier. The effectiveness of the approach is assessed by the Case Western Reserve University dataset: the analysis demonstrates the proposed method’s capabilities in accurately diagnosing faults in rotating equipment as compared to existing approaches.

The increase of internet usage in recent times has been a noticeable change in this generation. Users from all over the world use social sites to interact across the world. Countless websites are present today. With countless networks and sites, some people or companies tend to create new ways to lure out the random users using the web, such as phishing. In phishing, the normal users are swindled to use the fraudulent websites. The aim is to identify the phishing websites with great accuracy and compare different methods by which phishing websites can be tracked in an easier and more accurate way. Comparative studies of various algorithms are tested with the help of 10,000 datasets, each tested with 18 different parameters to increase the accuracy score of each algorithm. The paper shows the methods used for phishing detection are more accurate than other practices done so far using certain appropriate parameters and more useful.

Artificial Intelligence (AI) is becoming pervasive in most engineering domains, and railway transport is no exception. However, due to the plethora of different new terms and meanings associated with them, there is a risk that railway practitioners, as several other categories, will get lost in those ambiguities and fuzzy boundaries, and hence fail to catch the real opportunities and potential of machine learning, artificial vision, and big data analytics, just to name a few of the most promising approaches connected to AI. The scope of this paper is to introduce the basic concepts and possible applications of AI to railway academics and practitioners. To that aim, this paper presents a structured taxonomy to guide researchers and practitioners to understand AI techniques, research fields, disciplines, and applications, both in general terms and in close connection with railway applications such as autonomous driving, maintenance, and traffic management. The important aspects of ethics and explainability of AI in railways are also introduced. The connection between AI concepts and railway subdomains has been supported by relevant research addressing existing and planned applications in order to provide some pointers to promising directions.

Abstract The current travel demand in railways requires the adoption of novel approaches and technologies in order to increase network capacity. Virtual Coupling is considered one of the most innovative solutions to increase railway capacity by drastically reducing train headway. The aim of this paper is to provide an approach to investigate the potential of Virtual Coupling in railways by composing stochastic activity networks model templates. The paper starts describing the Virtual Coupling paradigm with a focus on standard European railway traffic controllers. Based on stochastic activity network model templates, we provide an approach to perform quantitative evaluation of capacity increase in reference Virtual Coupling scenarios. The approach can be used to estimate system capacity over a modelled track portion, accounting for the scheduled service as well as possible failures. Due to its modularity, the approach can be extended towards the inclusion of safety model components. The contribution of this paper is a preliminary result of the PERFORMINGRAIL (PERformance-based Formal modelling and Optimal tRaffic Management for movING-block RAILway signalling) project funded by the European Shift2Rail Joint Undertaking.

Future smart factories will be increasingly required to predict expected performance and dependability metrics related to their production processes. Domain-specific metrics include overall equipment effectiveness that measures production system availability/uptime, performance/speed and output quality. In this work-in-progress paper, we take initial steps towards a model-based approach to evaluate production-specific metrics using domain-specific languages, model transformations and stochastic modelling formalism.

Railway infrastructure operators need to push their network capacity up to their limits in high-traffic corridors. Virtual coupling is considered among the most relevant innovations to be studied within the European Horizon 2020 Shift2Rail Joint Undertaking as it can drastically reduce headways and thus increase the line capacity by allowing to dynamically connect two or more trains in a single convoy. This paper provides a proof of concept of Virtual coupling by introducing a specific operating mode within the European rail traffic management system/European train control system (ERTMS/ETCS) standard specification, and by defining a coupling control algorithm accounting for time-varying delays affecting the communication links. To that aim, we define one ploy to enrich the ERTMS/ETCS with Virtual coupling without changing its working principles and we borrow a numerical analysis methodology used to study platooning in the automotive field. The numerical analysis is also provided to support the proof of concept with quantitative results in a case-study simulation scenario.

Cybersecurity threats have surged in the past decades. Experts agree that conventional security measures will soon not be enough to stop the propagation of more sophisticated and harmful cyberattacks. Recently, there has been a growing interest in mastering the complexity of cybersecurity by adopting methods borrowed from Artificial Intelligence (AI) in order to support automation. In this paper, we provide a brief survey and some hints about Bayesian Network applications to intelligent cybersecurity in order to enable quantitative threat assessment for superior risk analysis and situation awareness.

The security of railway and mass-transit systems is increasingly dependant on the effectiveness of integrated Security Management Systems (SMS), which are meant to detect threats and to provide operators with information required for alarm verification purposes. In order to lower the false alarm rate and improve the detection reliability of threat scenarios, event correlation capabilities need to be integrated into the SMS. In this paper an existing approach based on a-priori defined event patterns is extended using a heuristic situation recognition approach which is more robust to both imperfect scenario modeling (human faults) and missed detections (sensor faults). The approach is based on similarity analysis between the event trees representing scenarios and it is effective both on-line and off-line. Applied on-line, it allows for an earlier and more fault-tolerant threat detection, since scenario matching is not required to be complete nor exact. Applied off-line, its effectiveness is twofold: first, it allows for detecting redundancies when updating the scenario repository; secondly, it enhances the post-event forensic search of suspicious behaviors not previously stored in the scenario repository. The strategy is being experimented in the context of railway protection.

Abstract Computer‐based control systems have grown in size, complexity, distribution and criticality. In this paper a methodology is presented to perform an ‘abstract testing’ of such large control systems in an efficient way: an abstract test is specified directly from system functional requirements and has to be instantiated in more test runs to cover a specific configuration, comprising any number of control entities (sensors, actuators and logic processes). Such a process is usually performed by hand for each installation of the control system, requiring a considerable time effort and being an error‐prone verification activity. To automate a safe passage from abstract tests, related to the so‐called generic software application, to any specific installation, an algorithm is provided, starting from a reference architecture and a state‐based behavioural model of the control software. The presented approach has been applied to a railway interlocking system, demonstrating its feasibility and effectiveness in several years of testing experience. Copyright © 2008 John Wiley & Sons, Ltd.