Physics-informed anomaly detection for industrial control systems (ICSs) is usually validated on synthetic data. How much of that performance survives on real hardware-in-the-loop (HIL) data is an open question. We re-implement PhySec-Edge, a hybrid BiLSTM/β-VAE/GAT/Kalman pipeline, as a full architecture and evaluate it on HAI 23.05. The gap is consistent across five seeds. AI-only F1 falls from 0.774 (synthetic) to 0.525 (HAI); the full hybrid falls from 0.798 to 0.512. The components diverge. PVE residual sharing transfers but weakens on real data (ΔF1 = +0.010 ± 0.008 across five seeds against +0.017 synthetic; ΔMCC = +0.013, positive in every seed). The PVE NIS decision score fails outright (raw FPR = 0.482; a correction brings it to 0.013, but detection does not recover). Under the retrospective fixed-FPR comparison, the preferred configuration is GNN+BiLSTM (F1 = 0.530 ± 0.002), outperforming full-EADE fusion (0.512 ± 0.003). From this split we derive the information-vs-decision pathway distinction, a way to reason about component-level generalization in hybrid architectures. We then set out what it means for physics-informed ICS detector design and edge deployment.

Greater decentralization of modern IT environments, driven by distributed infrastructure, IoT devices, etc., has highlighted the limitations of traditional perimeter based security models, which are based on the assumption that the internal network is trusted, while threats exist exclusively outside of it. Implicit trust and broad internal access that perimeter environments have can create – and do create – favorable conditions for lateral movement, insider misuse, and unauthorized access. This paper examines Zero Trust Architecture (ZTA) as a security paradigm designed to address these weaknesses. In parallel, the paper analyzes Li-Fi (Light Fidelity) communication as an alternative wireless technology whose physical properties, including spatial signal limitation and reduced penetration through walls, offer significant cybersecurity advantages compared to conventional radio frequency systems, but since the physical environment alone does not guarantee protection, cryptographic protection mechanisms in Li-Fi environments are emphasized. By combining the Zero Trust model with Li-Fi communication and secure key management mechanisms, a layered approach is provided for securing sensitive internal wireless communication against eavesdropping, message manipulation, and unauthorized access.

: This paper presents an approach for improving the performance of TinyML models through metaheuristic optimization. Due to limited computational resources in IoT environments, achieving high classification accuracy remains a challenge. The proposed method utilizes Particle Swarm Optimization (PSO) for tuning convolutional neural network (CNN) hyperparameters in the task of plant disease classification. Experimental results indicate that optimized models outperform baseline configurations in terms of accuracy and stability, with an improvement of over 20% in MCC compared to the baseline model. This study represents a part of a broader doctoral research focused on efficient and explainable AI models for resource-constrained environments.

In an epoch wherein the significance of cybersecurity has reached paramount importance, Li-Fi (Light Fidelity) emerges as a highly promising wireless communication paradigm that leverages visible light for the purpose of data transmission. In contrast to Wi-Fi, which is predicated on radio frequencies, Li-Fi utilizes the modulation of LED light to facilitate high-speed, secure, and interference free connectivity. This manuscript investigates the operational principles of Li-Fi, its advantages, limitations, and prospective applications across diverse domains, including healthcare, transportation, and industrial settings. Notable advantages encompass augmented security, enhanced efficiency, and diminished electromagnetic interference, rendering Li-Fi a viable alternative in sensitive contexts such as medical facilities and aviation environments. Nonetheless, obstacles such as restricted range and reliance on light sources impede its extensive adoption. Furthermore, the manuscript addresses fraud prevention strategies in digital transactions, underscoring the pivotal role of SEON in the identification and alleviation of cyber threats. By amalgamating Li-Fi with advanced cybersecurity frameworks, the research accentuates innovative solutions for secure and efficient digital communication. The conclusions indicate that Li-Fi, in conjunction with robust cybersecurity infrastructures, has the potential to markedly improve the security and efficiency of data transmission within the continuously evolving digital milieu.

Li-Fi technology represents a revolutionary approach to wireless communications, enabling high data transmission speeds, low latency, and enhanced security compared to traditional Wi-Fi and RF networks. Its integration with the Internet of Things (IoT) and Artificial Intelligence (AI) opens up new possibilities across various industries, including smart cities, industrial automation, and healthcare. Research has shown that Li-Fi can improve the efficiency of IoT applications through better resource management, faster data transmission, and increased communication security. In industrial environments, the use of Li-Fi in manufacturing has resulted in a 1.52% increase in efficiency and a reduction in errors, confirming its potential in real-world scenarios. Additionally, AI algorithms combined with Li-Fi technology enable automated decision-making, real-time data analysis, and process optimization. The development of advanced handover algorithms contributes to communication stability in multi-user environments, ensuring greater system reliability. Li-Fi technology has the potential to become the standard for future wireless communications, providing solutions to challenges in the IoT ecosystem. Its implementation will facilitate the development of smart and energy-efficient networks, accelerating the digital transformation of various sectors.

This paper explores the transformative impact of digital technologies on the fitness industry, focusing specifically on the role of artificial intelligence (AI) in enhancing modern exercise equipment. By analyzing the needs of recreational users and professional athletes, it examines how AI-driven fitness devices optimize personalization, improve performance tracking, and elevate the overall user experience. Using G&S Premium Gym in Bijeljina—the first fitness center in Bosnia and Herzegovina to integrate AI-powered equipment—as a case study, the paper delves into the technical specifications, functionality, and user interaction with these advanced machines. Key findings reveal that AI technologies significantly enhance training efficiency and customization, contributing to measurable improvements in user satisfaction and physical performance. However, challenges persist, particularly regarding technology accessibility, user digital literacy, and data privacy concerns. The research highlights the potential for AI to redefine standards in recreation while addressing these challenges. Recommendations for future research and implementation emphasize the importance of affordable, user-friendly AI solutions and improved data security measures.

Modern lifestyle, the abundance of high-calorie food, and the lack of physical activity have led to a global increase in obesity rates. In this context, the development of smart fitness equipment powered by artificial intelligence (AI) represents significant potential for enhancing physical activity, monitoring health parameters, and providing personalized guidance for weight loss and fitness improvement. This paper explores the functionalities and advantages of AI-equipped fitness devices, including physiological parameter monitoring, motion analysis through computer vision, detection of improper exercise execution, and automated adjustment of training plans. Special attention is given to the role of these technologies in obesity prevention among adults, as well as their potential for early detection of health risks. In addition to the literature review, the paper incorporates a practical case study conducted in Bijeljina, Bosnia and Herzegovina, where AI-integrated smart fitness equipment is already in use. This example illustrates the real-world implementation of such technologies and highlights their relevance for improving public health outcomes. Through a combined methodological approach, the paper examines actual and practical examples of successful applications of AI-powered fitness systems and discusses the challenges of their broader implementation within the context of public health.

: This paper presents a comprehensive study on the development of a traffic sign recognition system based on computer vision, which is crucial for improving road safety and supporting intelligent transport systems (ITS). By using advanced image processing techniques and machine learning algorithms, this research aims to create a reliable system capable of accurately recognizing and classifying various traffic signs. The methodology includes collecting a diverse set of traffic sign images, applying image enhancement techniques, and utilizing deep learning models for precise recognition in different environmental conditions. Experimental results indicate the system’s high accuracy in recognizing traffic signs, even in cases with complex backgrounds, demonstrating its potential for integration into autonomous vehicles and advanced driver assistance systems (ADAS). This research contributes to efforts aimed at enhancing road safety and improving traffic management through automated traffic sign recognition.

This paper explores the integration of computer vision technologies to enhance traffic safety through the effective detection of car-pedestrian interactions. As urban environments become more congested, pedestrian safety remains a critical concern. The system’s performance was evaluated using real-life footage from vehicle-mounted cameras, as well as images and videos sourced from online platforms. These real-world scenarios enabled a detailed assessment of the system’s accuracy and efficiency in practical conditions. The study highlights the potential for significant improvements in traffic safety, particularly in Bosnia and Herzegovina, where over 38% of registered vehicles are older than 23 years, and nearly 62% exceed 14 years. The aging vehicle fleet heightens the risk of accidents, underscoring the need for advanced detection methods. The proposed system automates the identification of hazardous situations on roads, allowing timely responses from relevant authorities.