https://www.mdu.se/

The primary purpose of this paper is to represent a structured comprehensive guide for doctoral candidates focusing on innovation and excellence in the rail sector. The work, developed in the framework of the PhDEU Rail project, funded by EU Rail Joint Undertaking, is designed to support candidates in aligning their work with the goals of Europe’s rail research initiatives and the future mobility systems. It aims to facilitate the understanding of the PhD process, combining requirements and outcomes with distinctive nature of rail research-oriented doctoral pathways. Helping candidates to align their research with the priorities of European rail system, promoting sustainable. The alignment of the PhD research with efficient and integrated transport solutions, including cutting-edge technologies, practices and research competencies, requires comprehensive literature reviews and rigorous methodologies. Additionally, building networks, engaging collaborative relationships with stakeholders can maximize impacts and empower candidates with knowledge, skills, and connections useful for their studies and future careers.

This paper provides an overview of the main results achieved within the Horizon 2020 Shift2Rail project named RAILS (Roadmaps for Artificial Intelligence Integration in the Rail Sector). The RAILS roadmapping process provided state-of-the-art, taxonomy, future research directions, and recommendations in three macro areas: Railway Safety and Automation, Predictive Maintenance and Defect Detection, and Traffic Planning and Management. RAILS findings shed light on the potential of intelligent technologies and provided essential guidelines for integrating machine learning into next-generation smart railways. 

Smart-troubleshooting refers to the semiautomatic process that identifies system failures, matches them with relevant troubleshooting information, and applies recovery or reconfiguration actions. The concept of smarttroubleshooting originates from the necessity to manage the increasing complexity of interconnected devices in environments like human robot interaction, smart cities, Internet of Things, Industry 4.0, and cyber-physical systems. Smart-troubleshooting remains incomplete, with existing studies addressing only parts of the problem. We propose a methodology that integrates product information and log files, enabling data collection, analysis, and synthesis to enhance system resilience and enable self-healing. First, we describe the methodology that implements the vision of smart-troubleshooting by integrating product information and log files to automate failure detection and resolution in interconnected systems. The methodology introduces key innovations by automating the correlation of failures through the integration of product data and log files, detecting real-time patterns using data mining techniques, and incorporating a feedback loop that allows the system to continuously improve its response to future issues. We validate the methodology through a mixed methods approach, using surveys and interviews to gather expert feedback and assess its practical applicability. In particular, we distributed the survey to 40 practitioners in the maintenance domain. We extracted, analyzed, and synthesized the survey data, complementing it with qualitative insights from four online, semi-structured, in-depth interviews. 

In Industry 4.0, efficient fault diagnosis is crucial for predictive maintenance but is often hindered by significant domain shifts between training and testing domains and lack of training datasets, limiting the effectiveness of machine learning in practice. Transfer learning has been used to address these challenges by utilizing knowledge from similar source domains. However, the scarcity of faulty data from real machines and the difficulty of obtaining labeled datasets from lab machines as source domain pose significant obstacles. This paper presents a novel diagnostic framework that integrates digital twins and transfer learning to overcome these limitations. Digital twins generate training datasets, while a model update strategy based on parameter sensitivity analysis improves adaptability. The framework also incorporates a partial transfer diagnostic model with a double attention mechanism to handle data distribution discrepancies and label inconsistencies between digital twins and real machines. Validated on an industrial rotating machine case study using real data, the proposed approach improves diagnostic accuracy by over 10% compared to state-of-the-art methods.

In recent times, the swift advancements in Artificial Intelligence (AI), like the ability of Artificial Intelligence to create systems capable of acquiring knowledge through experience, and competency to adjust for novel circumstances, enhance their efficacy as time progresses, and also AI mainly used developed systems that can reason and make decisions under uncertainty and Blockchain technology that involves the storage of data in a series of blocks that are interconnected where in each block encompasses a cryptographic hash of the preceding block, thereby rendering any attempts to manipulate the data exceedingly challenging and Distributed ledger technology, which facilitates secure, transparent, and unalterable transactions, is a prominent innovation in Blockchain. These two significant technologies (Artificial Intelligence (AI) and Blockchain) have presented novel opportunities for various industries. Among the numerous domains where these technologies could have a substantial influence lies the defense and military operations field. As countries across the globe endeavor to uphold their security and safeguard their interests, the incorporation of AI and blockchain into defense strategies can bring about a revolutionary transformation in how military operations are executed, thereby ensuring more protection in years to come. This paper provides a comprehensive study of various applications that use Artificial Intelligence (AI) and blockchain that are used in present-day military services and the future.

Significant opportunities and challenges, particularly in relation to cybersecurity, have been brought about by the growth of e-commerce. The research on cybersecurity concerns in e-commerce is summarized in this review, which also highlights major threats like malware, Distributed Denial-of-Service attacks, social engineering, and data breaches. These risks jeopardize user privacy and reveal weaknesses in online platforms, particularly in developing nations where trust, security, and technology infrastructure are barriers to adoption.Although they face implementation and scalability challenges, emerging technologies like blockchain provide enhanced security and transparency for e-commerce transactions. To combat increasingly complex cyberattacks, tactics like multifactor authentication, Artificial Intelligence–driven threat detection, and secure data storage are essential. Due to their limited cybersecurity resources, research also shows that small and medium-sized businesses (SMBs) are especially vulnerable. Research also shows that SMBs frequently do not have enough protection, highlighting a vulnerability.Strong legal frameworks, privacy regulations, and industry cooperation are necessary to lessen these risks. Improving e-commerce cybersecurity necessitates a multifaceted strategy that includes customer trust, organizational preparedness, technological advancements, and extensive policy frameworks. The need for customized solutions and comprehensive security frameworks to safeguard various market contexts is highlighted by this review.In order to enhance the security posture of e-commerce platforms, future research ought to broaden quantitative analyses, create international standards, and encourage cooperation across national borders. Businesses can improve resilience against cyber threats and navigate the complex cybersecurity landscape more skillfully by filling in these gaps. © 2026 selection and editorial matter, Gururaj H L, B Ramesh, Chandrika J and Hong Lin; individual chapters, the contributors.

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 near-random 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. © 2025 Copyright held by the owner/author(s).

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 compare FedAvg and 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.

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.