Deep auto-encoders (AEs) are widely employed deep learning methods in the field of anomaly detection across diverse domains (e.g., cybersecurity analysts managing large volumes of alerts, or medical practitioners monitoring irregular patient signals). In such contexts, practitioners often face challenges of scale and limited processing resources. To cope, strategies such as false positive reduction, human-in-the-loop review, and alert prioritization are commonly adopted. This paper explores the integration of uncertainty quantification (UQ) methods into alert prioritization for anomaly detection using ensembles of AEs. UQ models highlight doubtful classification decisions, enabling analysts to address the most certain alerts first, since higher certainty typically correlates with greater accuracy. Our study reveals a nuanced issue where applying UQ to ensembles of AEs can produce skewed distributions of large reconstruction errors (errors exceeding a pre-defined threshold), which may falsely suggest high uncertainty when standard deviation is used as the metric. Conventionally, a high standard deviation indicates high uncertainty. However, contrary to intuition, large reconstruction errors often reflect AE is strongly confident that an input is anomalous—not uncertainty about it. Moreover, ensembles of AEs generate reconstruction errors with varying ranges, complicating interpretation. To address this, we propose an extension that calibrates the standard deviation distribution of uncertainties, mitigating erroneous prioritization. Evaluation on 10 benchmark datasets demonstrates that our calibration approach improves the effectiveness of UQ methods in prioritizing alerts, while maintaining favorable trade-offs across other key performance metrics.

This paper presents the results of a cybersecurity audit conducted on thirty industrial SMEs located in a remote region of Eastern Canada. These firms face growing cyber threats while having limited access to security expertise and infrastructure. Using a mixed-method approach combining on-site technical assessments, structured interviews, and questionnaires, the study analyzes vulnerabilities through the TOE framework (Technological, Organizational, Environmental). Results show that 90% of companies lacked internal network segmentation, 80% were vulnerable to phishing attacks, and over 70% had no cybersecurity training or formal security policy. Based on these findings, we propose a set of low-cost and practical recommendations tailored to SMEs in isolated regions. These include awareness training, simple network protections, and internal policy development. The study highlights the urgent need for targeted cybersecurity strategies adapted to geographic and resource constraints, and contributes to both academic and operational understanding of how to improve cyber resilience in decentralized industrial ecosystems.

Anomaly detection in Open Charge Point Protocol (OCPP) is key to securing the infrastructure of electric vehicle charging stations (EVCS), which are increasingly vulnerable to cyber threats. Anomaly detection enables cyber-attack detection, and supervised learning is commonly used. However, its dependence on labeled datasets, often sparse and unbalanced, limits its detection ability on advanced threats. This study highlights the limitations of supervised learning techniques and explores clustering in semi-supervised and unsupervised learning techniques. To our knowledge, no previous work focuses on anomaly detection using clusters for semi-supervised and unsupervised learning applied to the CICEVSE2024 dataset. Our key contribution enhances OCPP security, enabling more robust anomaly detection in EVCS infrastructure and the broader smart grid ecosystem.

Intrusion Detection Systems (IDS) are essential for Internet of Things (IoT) security, but single models often fail due to IoT data heterogeneity. While machine learning ensembles combine complementary strengths, conventional static weighting schemes, such as majority voting and temporal stacking, do not adapt to sample-specific features and may underperform in diverse IoT scenarios. To address these limitations, we propose a dynamic feature-weighting ensemble framework for intrusion detection in IoT networks that combines adaptive weighting with a selected set of complementary base models suited to different traffic patterns. The approach combines four complementary models: Gradient Boosted Trees (LightGBM), Bagging-based Random Forest (RF), Instance-based k-Nearest Neighbors (kNN), and Deep Feedforward Neural Networks (FNN). It dynamically adjusts their weights based on the active features of each incoming traffic flow, emphasizing models best suited to specific patterns (e.g., LightGBM for packet-header patterns, FNN for nonlinear TCP flag interactions). Evaluated on the CICIoT2023 dataset, the framework achieved 99.95% precision and 98.59% recall, resulting in a$73 \%-97 \%$reduction in false positives (FP) and a$9 \%-30 \%$reduction in false negatives (FN) compared to individual models.

Secure authentication remains a critical challenge in healthcare IoT (H-IoT) systems, where constrained devices must ensure data integrity, privacy, and resilience despite limited resources. This paper proposes EHKEA, a lightweight mutual authentication and key establishment protocol designed specifically for H-IoT environments. EHKEA relies solely on symmetric cryptographic primitives and ephemeral randomness to provide mutual authentication, forward secrecy and resistance to common attacks such as replay, impersonation, and man-in-the-middle intrusions. We formally verify EHKEA in the Tamarin prover under the Dolev-Yao adversary model, proving key security properties including injective agreement and session key secrecy. A detailed informal analysis further confirms its robustness against desynchronization, insider threats, and key compromise impersonation. Comparative analysis with recent H-IoT protocols demonstrates that EHKEA achieves superior efficiency while offering stronger security guarantees, making it well-suited for deployment in real-time healthcare monitoring applications.

This research explores trust management in IoT water treatment plants, emphasizing the collaboration between advanced technologies and qualified personnel. We measure trust based on actor-specific criteria by integrating tailored policy languages. Our approach stands out for humanizing trust, making it accessible and understandable. The methodology includes an in-depth literature analysis, identification of trust criteria, and innovative integration of policy languages, with positive validation in simulated environments. The results demonstrate significant improvements, highlighting a positive impact on security and collaboration. In conclusion, our innovative approach promises to reshape trust management, providing a comprehensive and user-friendly perspective in technologically complex environments.

Underground mine galleries can be considered as complex transmission lines where multipath, attenuation, reflection, diffraction and scattering effects are dominants. However, some companies have started to deploy modern wireless system networks in mine galleries with the objective of increasing safety and productivity. In the last decade, ultra-wideband (UWB) technology has gained much interest for its application to wireless communications. This paper reports on experimental results of UWB channel propagation in an underground mine. Eigen-decomposition and subspace-based statistical signal processing on the autocorrelation matrix of the channel impulse response are used. We apply information theoretic criteria to estimate the number of significant eigenvalues. This result is then used to calculate the RMS delay spread of the channel.

The necessity for wireless communications in underground mines is well understood. Some companies have started to deploy modern wireless system networks in mine galleries with the objective of increasing safety and productivity. In the last decade, ultra-wideband (UWB) technology has gained much interest for its applications in wireless communications. A number of UWB channel measurement have been published in the literature. However, all these works treated environments such as office buildings, residential or industrial. This paper reports on experimental results of UWB channel characterization in underground mines. The communication channel is still not well modelled in these environments. Important channel parameters such as path loss exponent, shadow fading, spatial correlation, small-scale fading, RMS delay spread and mean excess delay are investigated. This work has been carried out at the underground communications research laboratory LRCS, and at the experimental mine CANMET (Canadian Center for Minerals and Energy Technology) in Val-d'Or, Canada.

A procedure of measurement and analysis for the UWB channel in underground mines is presented. The measured data is saved in the frequency-domain via a vector network analyzer (VNA). We first compare two methods for the analysis of the path loss dependence on frequency. Then, we present results from an autoregressive modeling technique. We show that a two pole model is sufficient to represent the characteristics of the UWB channel in underground mines. This work was carried out by the underground communications research laboratory LRCS1, and the CANMET (Canadian Center for Minerals and Energy Technology) experimental mine in Val-d'Or, Canada.