This paper presents an analysis of the challenges encountered by visually impaired and deaf-blind individuals when navigating urban environments and utilizing public transportation systems. Additionally, data was collected regarding the knowledge and role of assistive technologies in daily life. Through the implementation of surveys, interviews, and analysis of existing solutions, key issues were identified in navigating the transportation network, obtaining environmental information, and addressing inadequately adapted information. The research proposes the AssistMove service, designed to deliver information based on defined user needs and the specific type of user impairment. Furthermore, the paper delineates the optimal information and communication technology required for information delivery using an appropriate device capable of displaying information via a Braille display. The AssistMove service employs deep learning algorithms for object detection, sensor technology, and cloud infrastructure for data processing. The paper places particular emphasis on the education of users and professionals to ensure the proper and effective application of assistive technology.
Cybersecurity is a subset of systems for the security and safety of all peoples. As with systems for public safety it, too, needs acceptance and participation of the peoples of a democratic society. In a world deeply in need of democratic cybersecurity, the Internet of Things (IoT) demonstrates the deepest need with its crazily distributed network of sensors, actuators and communications. We discuss legal and policy issues that intertwine with the technology to insure public security under law of IoT and the people it impacts.
Problems in combinatorial optimization, such as MaxCut, are particularly difficult for classical algorithms due to having exponential solution spaces. Classical algorithms specifically struggle with problems like MaxCut in combinatorial optimization due to the exponential space of solutions. Both classical brute-force methods and the Quantum Approximate Optimization Algorithm (QAOA) were applied and evaluated on a 3-node graph, with QAOA tested at levels ( p = 1, 2, 5 ). From the results, we observe that QAOA fully identified all optimal bitstrings, and its effectiveness increased with greater circuit depth. The findings indicate QAOA achieved all optimal bitstrings, completely driving performance increasing with additional circuit depth. These findings indicate that QAOA has the potential to solve small-scale combinatorial problems and offers a foundation for further research on scalability and physical implementation on quantum hardware.
This article focuses on the issue of phishing attack detection, one of the most widespread types of cyber attacks, resulting in the acquisition of sensitive data. Such attacks can, in the worst case, lead to identity theft, financial gain for the attacker, or disclosure of acquired data with the intent to damage reputation, among other consequences. This thesis, therefore, explores detecting this threat through a modern approach - machine learning. The practical part of the work involves comparing divisions of machine learning methods in evaluating efficiency of the models within these groups. Specifically, it compares predictive outcomes from supervised learning, unsupervised learning, and reinforcement learning models. The thesis concludes with an assessment of which models are most effective for predicting phishing attacks. The novelty of this research lies in the comparative analysis of multiple learning paradigms, including underexplored models like Perceptron, within a unified preprocessing pipeline, which provides new insights into their practical applicability.
Quantum key distribution networks (QKDNs) enable information-theoretic secure key exchange by leveraging principles of quantum mechanics to detect eavesdropping, making them resistant to both classical and quantum computational attacks. Unlike classical cryptography, QKD security does not rely on computational assumptions but on the physical properties of quantum states, ensuring that any eavesdropping attempt introduces detectable disturbances. This paper provides an overview of security threats in QKDNs, addressing both quantum hacking attacks targeting quantum components and classical attacks affecting classical network infrastructure elements. By reviewing currently available research and standardization efforts done by standardization bodies such as ETSI, ITU-T and ISO/IEC in the aspect of quantum key distribution network security, this research emphasizes the need for a holistic security assessment methodology for QKDNs and proposes a conceptual, holistic QKDN security assessment methodology that is applicable to a wide range of QKDN architecture and protocol types. This methodology consists of six phases which include asset identification, threat modeling, vulnerability analysis, security testing, risk assessment, and reporting and mitigation. This methodology will guide future research focused on identifying and developing tools, techniques, and procedures for each assessment phase.
This paper explores the human-centric vision of Industry 5.0, emphasizing its potential to advance sustainability, well-being, and work-life balance in the digital age. Building upon the technological advancements of Industry 4.0, Industry 5.0 shifts the focus toward human-machine collaboration, environmental responsibility, and employee empowerment. Key areas of discussion include the paradigm shift from automation-driven systems to human-centric innovation, the integration of sustainable practices in industrial processes, and strategies to enhance employee well-being in modern workplaces. By examining synergies between sustainability, human-centricity, and well-being, the paper highlights the transformative potential of Industry 5.0 to create resilient, inclusive, and environmentally conscious industrial ecosystems. Challenges and future opportunities for implementing Industry 5.0 principles are also addressed, providing a roadmap for aligning technology with human and societal needs. The originality of this study lies in connecting the human-centric vision of Industry 5.0 with a specific focus on well-being and work-life balance, which are often overlooked in the current literature. This work offers a new perspective on how Industry 5.0, through the synergy of technology and human needs, can contribute to creating sustainable and inclusive work ecosystems.
Incentivizing sustainable initiatives through tokenization is an innovative approach to reducing emissions from transportation by rewarding participants. Blockchain technology allows the creation of tokens that can be used to reward users who practice sustainable transportation options, such as public transportation, bicycles, or electric vehicles. Tokens can be used for discounts, benefits, or as part of a broader economic incentive to promote sustainable practices. We use the blockchain chain to lock in the value of alternatives. The research results show that blockchain blocks have the following values: Bicycle block (0.71); Electric vehicle block (0.76) and Public transport block (0.66). According to the research results, the electric car turned out to be the best alternative (0.76), that is, it has the highest value when using subsidies. The other two alternatives Bicycle (0.71) and Public transport (0.66) also allow the use of subsidies, but to a lesser extent. Public transport, although the cheapest, is ranked lowest due to its lower emissions and efficiency scores. The results also highlight the importance of balancing emissions, costs and efficiency when choosing environmentally friendly means of transport, as well as the importance of different criteria weights in the decision-making process for sustainable transport.
The assessment of acquired knowledge has always been one of the fundamental indicators of the effectiveness of the educational process. In other words, we do not know if the outcome of electronic testing is a result of knowledge or of guessing. The primary subjective of this study is to investigate the potential of facial-expressions and other human-computer interaction (HCI) parameters in improving the reliability of high-stakes electronic testing. For this purpose, we employed the Py-Feat Python toolbox to extract statistical features of seven core emotions from video recordings of 25 student-participants during electronic testing. Using Weka, we evaluated the collected dataset via classification, clustering and regression techniques. Although Random Forest classifier peaked at 69.1
In this study, we extend our previous research on fuzzy logic-based video quality assessment by incorporating a genetic algorithm (GA) to optimize fuzzy membership functions. Accurate video quality assessment is crucial for multimedia applications such as video streaming, gaming, and remote communication. However, conventional metrics often fail to capture the perceptual nuances of high frame rate content, particularly regarding motion smoothness, compression artifacts, and content complexity. Our existing fuzzy logic model estimates the Mean Opinion Score (MOS) by integrating key video parameters, namely, frame rate, compression level, spatial information, and temporal information to capture non-linear interactions. To address this, we propose a novel GA-enhanced approach that fine-tunes membership function parameters, improving the model’s alignment with subjective MOS ratings. Experimental results show that the GA-enhanced model yields higher Spearman’s rank-order and Pearson correlation coefficients compared to both the original model and traditional metrics.
Quantum Key Distribution (QKD) networks enable unconditionally secure key exchange using quantum mechanical principles; however, routing cryptographic keys across multi-hop quantum networks introduces challenges unique to quantum communication. This survey analyzes and classifies 26 routing strategies proposed between 2013 and 2024 for terrestrial, satellite, and hybrid QKD infrastructures. Dynamic, key-aware routing algorithms have been shown to reduce service rejection rates by 25–40
Stock trend forecasting, a challenging problem in the financial domain, involves extensive data and related indicators. Relying solely on empirical analysis often yields unsustainable and ineffective results. Machine learning researchers have demonstrated that the application of random forest algorithm can enhance predictions in this context, playing a crucial auxiliary role in forecasting stock trends. This study introduces a new approach to stock market prediction by integrating sentiment analysis using FinGPT generative AI model with the traditional Random Forest model. The proposed technique aims to optimize the accuracy of stock price forecasts by leveraging the nuanced understanding of financial sentiments provided by FinGPT. We present a new methodology called “Sentiment-Augmented Random Forest” (SARF), which incorporates sentiment features into the Random Forest framework. Our experiments demonstrate that SARF outperforms conventional Random Forest and LSTM models with an average accuracy improvement of 9.23
This research paper investigates the integration of Artificial Intelligence (AI), with a focus on Machine Learning (ML) and Deep Learning (DL), for bolstering cybersecurity defences against phishing attacks. Utilizing a comprehensive dataset of URL features, the study assesses the efficacy of various ML algorithms—namely Decision Tree, Logistic Regression, Support Vector Machine, Random Forest, and K-Nearest Neighbors—in pinpointing phishing websites. Research paper is conducted using Google Colaboratory, Python libraries and Weka tool. The research identifies the Random Forest algorithm as the most effective, demonstrating superior accuracy in detecting phishing URLs during both training and testing phases. The findings accentuate the pivotal role of AI in advancing cybersecurity measures, advocating for the incorporation of sophisticated AI technologies in the fight against cyber threats. Additionally, it outlines future research directions, including the enhancement of model precision through the integration of more comprehensive data attributes. This paper significantly contributes to the cybersecurity and AI domains by showcasing the practical applications and benefits of AI in identifying and mitigating cyber risks.
This paper aims to investigate the integration of electromobility into public transport systems through a case study. Given the growing global environmental concerns and the need to reduce greenhouse gas emissions, electromobility is becoming a key factor in the transition to more sustainable forms of transport. This paper analyses the strategic, technical and social aspects of integrating electric vehicles into existing public transport systems. The case study highlights the challenges, opportunities and best practices associated with this integration through a comprehensive approach that includes quantitative and qualitative methodologies. Emphasis is placed on the analysis of infrastructure needs, such as charging stations, fleet modifications and intelligent control systems, which are essential for the effective integration of electromobility into public transport systems. Furthermore, the paper assesses the impact of electromobility. The case study findings provide important learning and recommendations for city planners, policymakers and public transport operators seeking to implement or expand electromobility in their systems. This paper contributes to the growing body of research on electromobility in the context of sustainable mobility. It offers valuable insights for the future development and integration of electromobility into public transport at a global level.
Our work presents a novel approach to creating new encryption algorithms. We were inspired by genetic algorithms based on the Darwin evolution theory. As the method of invention, we chose the crossover technique. We aim to create unique encryption algorithms that will be more secure, faster, and require less memory space or computational energy. New encryption algorithms are based on the most commonly used encryption techniques. During the crossover process, we consider the most critical metrics for encryption algorithms. We also briefly described the crossover and mutation process. The developed method is in the state of design. The design presented in this article is in the form of a flowchart diagram and detailed decryption. We plan to implement it and test it.
The paper explores the performance dynamics of a tandem queueing system (TQS) comprising two interconnected nodes and assesses its responsiveness under varying service conditions. By using simulation tools, we conducted a range of tests to discover the variability of waiting times in the queues and the system, for different server utilization, as well as for cases when the service times follow exponential and normal distributions. We discovered the interplay between the nodes, i.e., how the performance of one node affects the other within the same TQS. Namely, we showed how the waiting times in the queues are crucial in understanding the TQS dynamics. We also compared the simulation results with the well-established analytical models and discovered that their applicability is limited. Particularly noteworthy was the disparity observed in the performance estimation of the second node within the TQS across all simulations, indicating potential over- or underestimation by analytical models.
This paper introduces a novel image dataset tailored for evaluating machine learning solutions, particularly focusing on deep neural networks. Derived from X-ray images of wheat grains, the dataset encompasses three distinct species: Kama, Rosa, and Canadian. We provide a comprehensive overview of the dataset’s structure and conduct experiments using ten pretrained deep neural networks to classify wheat species. The Seeds Image Data Set offers a competitive alternative to established object recognition benchmarks such as CIFAR-10, CIFAR-100, SVHN, and ImageNet. Its compact size streamlines computational processes, making it an efficient resource for exploratory data analysis. The dataset will be publicly available, serving as a foundational resource for future research endeavors in the field.
ICT - Information and communications technology as an important stakeholder and driver of modern progress. Since the investments in ICT are significant, there are also high and sometimes unrealistic expectations in terms of optimizing business quality and achieving a higher level of efficiency. All these expectations require high-quality and efficient management. The most crucial part of managing ICT in any organization is making good and timely decisions that will create new value and consistently, an expected competitive advantage. This is where the CIO - Chief Information Officer becomes an important factor because his skills and knowledge can optimize the organization. Otherwise, in the absence of the necessary knowledge and skills, the CIO can also set the organization back. Important decisions are made at higher hierarchical levels of the organization, so it is very important to emphasize that the CIO should also participate in the decision-making process. To participate in such decisions, it is crucial for the CIO to become a member of TMT - top management team. If the management has an inadequate level of ICT literacy, then it is questionable whether the CIO can even get the opportunity to optimize something, because there is a possibility that the expectations and understanding of the importance of the CIO are unrealistic. CIOs must also acquire the necessary managerial and communication skills to increase their influence. In this way, they will not only secure a board seat but also be able to participate in all important decisions.
Increasing productivity is one of the primary goals in the context of economic sustainability. This will bring several challenges, which, however, can be managed through the digital technologies that the 21st century brings. The growth rate of digitization is seen in every industry. As an essential sector of the economy, the construction industry has the same plans. Objectives in circular economy and sustainability in the construction industry are also prioritized for the management of the industries. Innovative modelling of buildings, like the implementation of new digital technologies, also brings many challenges, but on the other hand, also opportunities. This process in construction project management can also be key to improving sustainable productivity. This research aims to map the current state of the use of information modelling of buildings and their impact on yield. The research was carried out on 199 construction projects in three countries (Slovakia, Croatia, and Slovenia). Data collection was ensured by questionnaire inquiry. Subsequently, the data were processed, and essential tools of statistical processing were used. In addition to descriptive statistics, data redistribution tests and subsequent retesting of statistical significance, namely ANOVA and the Kruskal-Wallis’s test, were used. Correlation analysis was used to examine the dependencies between variables. The survey results highlight the impact of building information modelling in construction project management for sustainable productivity improvements. Research has shown that information modeling of constructions brings an increase in productivity, including through a circular economy approach, when machines and equipment, technologies, and materials are used more efficiently, which leads to the optimization of costs.
This article deals with the use of Product Lifecycle Management (PLM) in the preparation of logistics process simulation. PLM is a comprehensive information system that provides management and tracking of the product lifecycle from the conceptual phase to disposal. Logistics process simulation is a tool that allows to analyze and optimize the course of logistics operations without the risk of real interventions. In this abstract, we present an example of the use of PLM in the preparation of a simulation of logistics processes in a manufacturing environment. We demonstrate how PLM allows us to model different logistics scenarios, optimize inventory, minimize time and financial costs, and increase overall supply chain efficiency. The combination of PLM and logistics process simulation opens the way to improving supply chain management and increasing business competitiveness through more accurate and efficient logistics solutions. This abstract discusses the benefits and potential of this combination and suggests further avenues for research and implementation in practice. The paper highlights the synergy between PLM and simulation technologies, illustrating how PLM acts as a bridge between disparate data sources, ensuring a unified and accurate representation of the product and its associated logistics processes. With this integration, organizations can create realistic simulations that reflect the complexity of their supply chains, enabling them to identify potential bottlenecks, optimize resource allocation, and increase overall operational resilience.
Urban environments, with an ever-increasing population and ever-increasing traffic intensity, today represent a challenge for research and implementation of systems that would improve all aspects of life in such environments. These systems, based on smart technologies, are increasingly present in urban environments. The application of modern information technologies is the basis of these systems, and the application of cloud-based systems, which are key to the integration of systems and data, is particularly important. This paper presents a model of a cloud-based system for air quality monitoring in an urban traffic environment, with the main focus on software services deployed in the cloud. In addition to software services in the cloud, organized in service-oriented architecture (SOA), other relevant parts of the system with which these services communicate are also shown: the sensing layer, local area maps online service, online meteorological service, and a variety of web and mobile user applications. The main feature of the model, which contributes to its application in various environments, is that it is platform-independent, scalable and flexible for possible adaptations depending on proposed requirements.
