Appl. Syst. Innov., Volume 8, Issue 4 (August 2025) – 27 articles

In this work we present the shape optimization of a 2000 L water tank placed behind the rear axle of a forestry skidder. The main criterion is the static stability of the vehicle. The purpose of the research is to decrease the impact of the tank on stability of the vehicle. The stability is determined in the form of distances of vectors of a stability triangle and a gravity vector. The tank is divided into small elements and their impact on stability is evaluated independently. Then, the gravity vector, placed in the center of gravity of the vehicle with the tank, combines the gravities of the vehicle and the tank composed of as many elements as required for the desired volume. The Python 3.13 programming language is used to implement the solution. The results for various shapes of the tank are displayed in the form of heatmaps. A slope angle of 20 degrees is used for the analysis. The results show that the longitudinal or lateral stability can be improved by shape modifications of the tank. The most interesting output is the final shape of the tank that improves terrain accessibility of the vehicle. The optimization method is universal and can also be used for different vehicles, tank placements and also auxiliary devices added in general positions. Full article

Glaucoma is a serious eye condition that damages the optic nerve and affects the transmission of visual information to the brain. It is the second leading cause of blindness worldwide. With deep learning, CAD systems have shown promising results in diagnosing glaucoma but mostly rely on small-labeled datasets. Annotated fundus image datasets improve deep learning predictions by aiding pattern identification but require extensive curation. In contrast, unlabeled fundus images are more accessible. The proposed method employs a semi-supervised learning approach to utilize both labeled and unlabeled data effectively. It follows traditional supervised training with the generation of pseudo-labels for unlabeled data, and incorporates self-supervised techniques that eliminate the need for manual annotation. It uses a twin self-supervised learning approach to improve glaucoma diagnosis by integrating pseudo-labels from one model into another self-supervised model for effective detection. The self-supervised patch-based exemplar CNN generates pseudo-labels in the first stage. These pseudo-labeled data, combined with labeled data, train a convolutional auto-encoder classification model in the second stage to identify glaucoma features. A support vector machine classifier handles the final classification of glaucoma in the model, achieving 98% accuracy and 0.98 AUC on the internal, same-source combined fundus image datasets. Also, the model maintains reasonably good generalization to the external (fully unseen) data, achieving AUC of 0.91 on the CRFO dataset and AUC of 0.87 on the Papilla dataset. These results demonstrate the method’s effectiveness, robustness, and adaptability in addressing limited labeled fundus data and aid in improved health and lifestyle. Full article

The inclusion of the Internet of Things (IoT) in indoor agricultural systems has become a fundamental tool for improving cultivation systems by providing key information for decision-making in pursuit of better performance. This article presents the design and implementation of an IoT-based agricultural system installed in a plant growth chamber for hydroponic cultivation under controlled conditions. The growth chamber is equipped with sensors for air temperature, relative humidity (RH), carbon dioxide (CO₂) and photosynthetically active photon flux, as well as control mechanisms such as humidifiers, full-spectrum Light Emitting Diode (LED) lamps, mini split air conditioner, pumps, a Wi-Fi surveillance camera, remote monitoring via a web application and three Nutrient Film Technique (NFT) hydroponic systems with a capacity of ten plants each. An ATmega2560 microcontroller manages the smart system using the MODBUS RS-485 communication protocol. To validate the proper functionality of the proposed system, a case study was conducted using lettuce crops, in which the impact of different nutrient solution concentrations (50%, 75% and 100%) on the phenotypic development and nutritional content of the plants was evaluated. The results obtained from the cultivation experiment, analyzed through analysis of variance (ANOVA), show that the treatment with 75% nutrient concentration provides an appropriate balance between resource use and nutritional quality, without affecting the chlorophyll content. This system represents a scalable and replicable alternative for protected agriculture. Full article

Target detection is essential in intelligent transportation and autonomous control of unmanned aerial vehicles (UAVs), with single-stage detection algorithms used widely due to their speed. However, these algorithms face limitations in detecting small targets, especially in aerial photography from unmanned aerial vehicles (UAVs), where small targets are often occluded, multi-scale semantic information is easily lost, and there is a trade-off between real-time processing and computational resources. Existing algorithms struggle to effectively extract multi-dimensional features and deep semantic information from images and to balance detection accuracy with model complexity. To address these limitations, we developed FFG-YOLO, a lightweight small-target detection method for UAVs based on YOLOv8. FFG-YOLO incorporates three modules: a feature enhancement block (FEB), a feature concat block (FCB), and a global context awareness block (GCAB). These modules strengthen feature extraction from small targets, resolve semantic bias in multi-scale feature fusion, and help differentiate small targets from complex backgrounds. We also improved the positioning accuracy of small targets using the Wasserstein distance loss function. Experiments showed that FFG-YOLO outperformed other algorithms, including YOLOv8n, in small-target detection due to its lightweight nature, meeting the stringent real-time performance and deployment requirements of UAVs. Full article

As voice cloning technology rapidly advances, the risk of personal voices being misused by malicious actors for fraud or other illegal activities has significantly increased, making the collection of speech data increasingly challenging. To address this issue, this study proposes a data augmentation method based on XText-to-Speech (XTTS) synthesis to tackle the challenges of small-sample, multi-class speech recognition, using profanity as a case study to achieve high-accuracy keyword recognition. Two models were therefore evaluated: a CNN model (Proposed-I) and a CNN-Transformer hybrid model (Proposed-II). Proposed-I leverages local feature extraction, improving accuracy on a real human speech (RHS) test set from 55.35% without augmentation to 80.36% with XTTS-enhanced data. Proposed-II integrates CNN’s local feature extraction with Transformer’s long-range dependency modeling, further boosting test set accuracy to 88.90% while reducing the parameter count by approximately 41%, significantly enhancing computational efficiency. Compared to a previously proposed incremental architecture, the Proposed-II model achieves an 8.49% higher accuracy while reducing parameters by about 98.81% and MACs by about 98.97%, demonstrating exceptional resource efficiency. By utilizing XTTS and public corpora to generate a novel keyword speech dataset, this study enhances sample diversity and reduces reliance on large-scale original speech data. Experimental analysis reveals that an optimal synthetic-to-real speech ratio of 1:5 significantly improves the overall system accuracy, effectively addressing data scarcity. Additionally, the Proposed-I and Proposed-II models achieve accuracies of 97.54% and 98.66%, respectively, in distinguishing real from synthetic speech, demonstrating their strong potential for speech security and anti-spoofing applications. Full article

This paper proposes a hybrid model that combines convolutional neural networks with natural language processing techniques for least significant bit-based steganography detection in grayscale digital images. The proposed approach identifies hidden messages by analyzing subtle alterations in the least significant bits and validates the linguistic coherence of the extracted content using a semantic filter implemented with spaCy. The system is trained and evaluated on datasets ranging from 5000 to 12,500 images per class, consistently using an 80% training and 20% validation partition. As a result, the model achieves a maximum accuracy and precision of 99.96%, outperforming recognized architectures such as Xu-Net, Yedroudj-Net, and SRNet. Unlike traditional methods, the model reduces false positives by discarding statistically suspicious but semantically incoherent outputs, which is essential in forensic contexts. Full article

The development of Industry 4-oriented facilities in academia for training and research purposes is playing a significant role in pushing forward the Fourth Industrial Revolution. This study can serve academic staff who are intending to build their Industry 4 facilities, to better understand the key features, constraints, and opportunities. This paper presents a systematic literature review of 145 peer-reviewed studies published between 2011 and 2023, which are identified across Scopus, SpringerLink, and Web of Science. As a result, we emphasise the significance of developing Industry 4 learning facilities in academia and outline the main design principles of the Industry 4 ecosystems. We also investigate and discuss the key Industry 4-related technologies that have been extensively used and represented in the reviewed literature, and summarise the challenges and roadblocks that current participants are facing. From these insights, we identify research gaps, outline technology mapping and maturity level, and propose a strategic roadmap for future implementation of Industry 4 facilities. The results of the research are expected to support current and future participants in increasing their awareness of the significance of the development, clarifying the research scope and objectives, and preparing them to deal with inherent complexity and skills issues. Full article

Entropy is a foundational concept across scientific domains, playing a role in understanding disorder, randomness, and uncertainty within systems. This study applies Shannon’s entropy in information theory to evaluate and manage complexity in industrial supply chain management. The purpose of the study is to propose a quantitative modeling method, employing Shannon’s entropy model as a proxy to assess the complexity in SCs. The underlying assumption is that information entropy serves as a proxy for the complexity of the SC. The research method is quantitative modeling, which is applied to four focal companies from the agrifood and metalworking industries in Southern Brazil. The results showed that companies prioritizing cost and quality exhibit lower complexity compared to those emphasizing flexibility and dependability. Additionally, information flows related to specially engineered products and deliveries show significant differences in average entropies, indicating that organizational complexities vary according to competitive priorities. The implications of this suggest that a focus on cost and quality in SCM may lead to lower complexity, in opposition to a focus on flexibility and dependability, influencing strategic decision making in industrial contexts. This research introduces the novel application of information entropy to assess and control complexity within industrial SCs. Future studies can explore and validate these insights, contributing to the evolving field of supply chain management. Full article

Social network platforms have a big impact on the development of companies by influencing clients’ behaviors and sentiments, which directly affect corporate reputations. Analyzing this feedback has become an essential component of business intelligence, supporting the improvement of long-term marketing strategies on a larger scale. The implementation of powerful sentiment analysis models requires a comprehensive and in-depth examination of each stage of the process. In this study, we present a new comparative approach for several feature extraction techniques, including TF-IDF, Word2Vec, FastText, and BERT embeddings. These methods are applied to three multilingual datasets collected from hotel review platforms in the tourism sector in English, French, and Arabic languages. Those datasets were preprocessed through cleaning, normalization, labeling, and balancing before being trained on various machine learning and deep learning algorithms. The effectiveness of each feature extraction method was evaluated using metrics such as accuracy, F1-score, precision, recall, ROC AUC curve, and a new metric that measures the execution time for generating word representations. Our extensive experiments demonstrate significant and excellent results, achieving accuracy rates of approximately 99% for the English dataset, 94% for the Arabic dataset, and 89% for the French dataset. These findings confirm the important impact of vectorization techniques on the performance of sentiment analysis models. They also highlight the important relationship between balanced datasets, effective feature extraction methods, and the choice of classification algorithms. So, this study aims to simplify the selection of feature extraction methods and appropriate classifiers for each language, thereby contributing to advancements in sentiment analysis. Full article

Public service innovation research often focuses on the private or general public sectors, leaving the distinct challenges government entities face unexplored. An empirical study was carried out to bridge this gap using survey results from the United Arab Emirates (UAE) government entities. This study built on that research by further analyzing the relationships among these challenges through a social network approach, visualizing and analyzing the connections between them by utilizing betweenness centrality and eigenvector centrality as key metrics. Based on this analysis, the challenges were classified into different categories; 8 out of 22 challenges were identified as critical due to their high values in both metrics. Addressing these critical challenges is expected to create a cascading impact, helping to resolve many others. Targeted strategies are proposed, and leveraging open innovation is highlighted as an effective and versatile solution to address and mitigate these challenges. This study is one of the few to adopt a social network analysis perspective to visualize and analyze the relationships among challenges, enabling the identification of critical ones. This research offers novel and valuable insights that could assist decision-makers in UAE government entities and countries with similar contexts with actionable strategies to advance public service innovation. Full article

Modern traffic management requires sophisticated approaches to address the complexities of urban road networks, which continue to grow in complexity due to increasing urbanization and vehicle usage. Traditional methods often fall short in mitigating congestion and optimizing traffic flow, inducing the exploration of innovative traffic control strategies based on advanced theoretical frameworks. In this sense, we explore different game theory-based control strategies in an eight-intersection traffic network modeled by means of hybrid systems and graph theory, using a software simulator that combines the multi-modal traffic simulation software VISSIM and MATLAB to integrate traffic network parameters and population game criteria. Across five distinct network scenarios with varying saturation conditions, we explore a fixed-time scheme of signaling by means of fictitious play dynamics and adaptive schemes, using dynamics such as Smith, replicator, Logit and Brown–Von Neumann–Nash (BNN). Results show better performance for Smith and replicator dynamics in terms of traffic parameters both for fixed and variable signaling times, with an interesting outcome of fictitious play over BNN and Logit. Full article

Optimization problems often feature local optima with a significant difference in the basin of attraction (BoA), making evolutionary computation methods prone to discarding solutions located in less-attractive BoAs, thereby posing challenges to the search for optima in these BoAs. To enhance the ability to find these optima, various niching methods have been proposed to restrict the competition scope of individuals to their specific neighborhoods. However, redundant searches in more-attractive BoAs as well as necessary searches in less-attractive BoAs can only be promoted simultaneously by these methods. To address this issue, we propose a general framework for niching methods named niching-driven divide-and-conquer hill exploration (NDDCHE). Through gradually learning BoAs from the search results of a niching method and dividing the problem into subproblems with a much smaller number of optima, NDDCHE aims to bring a more balanced distribution of searches in the BoAs of optima found so far, and thus enhance the niching method’s ability to find optima in less-attractive BoAs. Through experiments where niching methods with different categories of niching techniques are integrated with NDDCHE and tested on problems with significant differences in the size of the BoA, the effectiveness and the generalization ability of NDDCHE are proven. Full article

Modern manufacturing strives for peak efficiency while facing pressing demands for environmental sustainability. Balancing these often-conflicting objectives represents a fundamental trade-off in modern manufacturing, as traditional methods typically address them in isolation, leading to suboptimal outcomes. Process mining offers operational insights but often lacks dynamic environmental indicators, while standard Life Cycle Assessment (LCA) provides environmental evaluation but uses static data unsuitable for real-time optimization. Frameworks integrating real-time data for dynamic multi-objective optimization are scarce. This study proposes a comprehensive, data-driven, cloud-based framework that overcomes these limitations. It uniquely combines three key components: (1) real-time Process Mining for actual workflows and operational KPIs; (2) dynamic LCA using live sensor data for instance-level environmental impacts (energy, emissions, waste) and (3) Multi-Objective Optimization (NSGA-II) to identify Pareto-optimal solutions balancing efficiency and sustainability. TOPSIS assists decision-making by ranking these solutions. Validated using extensive real-world data from a tube manufacturing facility processing over 390,000 events, the framework demonstrated significant, quantifiable improvements. The optimization yielded a Pareto front of solutions that surpassed baseline performance (87% efficiency; 2007.5 kg CO₂/day). The optimal balanced solution identified by TOPSIS simultaneously increased operational efficiency by 5.1% and reduced carbon emissions by 12.4%. Further analysis quantified the efficiency-sustainability trade-offs and confirmed the framework’s adaptability to varying strategic priorities through sensitivity analysis. This research offers a validated framework for industrial applications that enables manufacturers to improve both operational efficiency and environmental sustainability in a unified manner, moving beyond the limitations of disconnected tools. The validated integrated framework provides a powerful, data-driven tool, recommended as a valuable approach for industrial applications seeking continuous improvement in both economic and environmental performance dimensions. Full article

Researchers conduct experiments to discover factors influencing the experimental subjects, so the experimental design is essential. The response surface methodology (RSM) is a special experimental design used to evaluate factors significantly affecting a process and determine the optimal conditions for different factors. The relationship between response values and influencing factors is mainly established using regression analysis techniques. These equations are then used to generate contour and surface response plots to provide researchers with further insights. The impact of regression techniques on response surface methodology (RSM) model building has not been studied in detail. This study uses complete regression techniques to analyze sixteen datasets from the literature on semiconductor manufacturing, steel materials, and nanomaterials. Whether each variable significantly affected the response value was assessed using backward elimination and a t-test. The complete regression techniques used in this study included considering the significant influencing variables of the model, testing for normality and constant variance, using predictive performance criteria, and examining influential data points. The results of this study revealed some problems with model building in RSM studies in the literature from three engineering fields, including the direct use of complete equations without statistical testing, deletion of variables with p-values above a preset value without further examination, existence of non-normality and non-constant variance conditions of the dataset without testing, and presence of some influential data points without examination. Researchers should strengthen training in regression techniques to enhance the RSM model-building process. Full article

In the context of Industry 4.0, Total Productive Maintenance (TPM) is undergoing a major shift driven by digital technologies such as the IoT, AI, cloud computing, and Cyber–Physical systems. This study explores how these technologies reshape traditional TPM pillars and practices through a two-phase methodology: bibliometric analysis, which reveals global research trends, key contributors, and emerging themes, and a systematic review, which discusses how core TPM practices are being transformed by advanced technologies. It also identifies key challenges of this transition, including data aggregation, a lack of skills, and resistance. However, despite the growing body of research on digital TPM, a major gap persists: the lack of a standardized model applicable across industries. Existing approaches are often fragmented or too context-specific, limiting scalability. Addressing this gap requires a structured approach that aligns technological advancements with TPM’s foundational principles. Taking a cue from these findings, this article formulates a systematic and scalable framework for TPM 4.0 deployment. The framework is based on four pillars: modular technological architecture, phased deployment, workforce integration, and standardized performance indicators. The ultimate goal is to provide a basis for a universal digital TPM standard that enhances the efficiency, resilience, and efficacy of smart maintenance systems. Full article

The accelerated development of Artificial Intelligence (AI) capabilities and systems is driving a paradigm shift in productivity, innovation and growth. Despite this generational opportunity, AI is fraught with significant challenges and risks. To address these challenges, responsible AI has emerged as a modus operandi that ensures protections while not stifling innovations. Responsible AI minimizes risks to people, society, and the environment. However, responsible AI principles and practice are impacted by ‘principle proliferation’ as they are diverse and distributed across the applications, stakeholders, risks, and downstream impact of AI systems. This article presents a systematic review of responsible AI principles and practice with the objectives of discovering the current state, the foundations and the need for responsible AI, followed by the principles of responsible AI, and translation of these principles into the responsible practice of AI. Starting with 22,711 relevant peer-reviewed articles from comprehensive bibliographic databases, the review filters through to 9700 at de-duplication, 5205 at abstract screening, 1230 at semantic screening and 553 at final full-text screening. The analysis of this final corpus is presented as six findings that contribute towards the increased understanding and informed implementation of responsible AI. Full article

Reinforced concrete deep beams (RCDBs) provide significant strength and serviceability for building structures. However, a simple, general, and universally accepted procedure for predicting their shear strength (SS) has yet to be established. This study proposes a novel data-driven approach to predicting the SS of RCDBs using an enhanced CatBoost (CB) model. For this purpose, a newly comprehensive database of RCDBs with shear failure, including 950 experimental specimens, was established and adopted. The model was developed through a customized procedure including feature selection, data preprocessing, hyperparameter tuning, and model evaluation. The CB model was further evaluated against three data-driven models (e.g., Random Forest, Extra Trees, and AdaBoost) as well as three prominent mechanics-driven models (e.g., ACI 318, CSA A23.3, and EU2). Finally, the SHAP algorithm was employed for interpretation to increase the model’s reliability. The results revealed that the CB model yielded a superior accuracy and outperformed all other models. In addition, the interpretation results showed similar trends between the CB model and mechanics-driven models. The geometric dimensions and concrete properties are the most influential input features on the SS, followed by reinforcement properties. In which the SS can be significantly improved by increasing beam width and concert strength, and by reducing shear span-to-depth ratio. Thus, the proposed interpretable data-driven model has a high potential to be an alternative approach for design practice in structural engineering. Full article

This paper addresses the problem of controlling quadcopters operating in an environment characterized by unpredictable disturbances such as wind gusts. From a control point of view, this is a nonstandard, highly challenging problem. Fundamentally, these quadcopters are high-order dynamical systems characterized by an under-actuated and highly nonlinear model with coupling between several state variables. The main objective of this work is to achieve a trajectory by tracking desired altitude and attitude. The problem was tackled using a robust control approach with a multi-loop nonlinear controller combined with extended Kalman filtering (EKF). Specifically, the flight control system consists of two regulation loops. The first one is an outer loop based on the backstepping approach and allows for control of the elevation as well as the yaw of the quadcopter, while the second one is the inner loop, which allows the maintenance of the desired attitude by adjusting the roll and pitch, whose references are generated by the outer loop through a standard PID, to limit the 2D trajectory to a desired set path. The investigation integrates EKF technique for sensor signal processing to increase measurements accuracy, hence improving robustness of the flight. The proposed control system was formally developed and experimentally validated through indoor tests using the well-known Parrot Mambo unmanned aerial vehicle (UAV). The obtained results show that the proposed flight control system is efficient and robust, making it suitable for advanced UAV navigation in dynamic scenarios with disturbances. Full article

Naive Bayes (NB) classifiers are widely used for their simplicity, computational efficiency, and interpretability. However, their predictive performance can degrade significantly in real-world settings where the conditional independence assumption is often violated. More complex NB variants address this issue but typically introduce structural complexity or require explicit dependency modeling, limiting their scalability and transparency. This study proposes two lightweight ensemble-based extensions—randomized feature naive Bayes (RF-NB) and randomized feature bootstrapped naive Bayes (RFB-NB)—designed to enhance robustness and predictive stability without altering the underlying NB model. By integrating randomized feature selection and bootstrap resampling, these methods implicitly reduce feature dependence and noise-induced variance. Evaluation across twenty real-world datasets spanning medical, financial, and industrial domains demonstrates that RFB-NB consistently outperformed classical NB, RF-NB, and k-nearest neighbor in several cases. Although random forest achieved higher average accuracy overall, RFB-NB demonstrated comparable accuracy with notably lower variance and improved predictive stability specifically in datasets characterized by high noise levels, large dimensionality, or significant class imbalance. These findings underscore the practical and complementary advantages of RFB-NB in challenging classification scenarios. Full article

Phishing attacks continue to pose serious challenges to cybersecurity, with attackers constantly refining their methods to bypass detection systems. One particularly evasive technique is Zero-Font phishing, which involves the insertion of invisible or zero-sized characters into email content to deceive both users and traditional email filters. Because these characters are not visible to human readers but still processed by email systems, they can be used to evade detection by traditional email filters, obscuring malicious intent in ways that bypass basic content inspection. This study introduces a proactive phishing detection tool capable of identifying both traditional and Zero-Font phishing attempts. The proposed tool leverages a multi-layered security framework, combining structural inspection and machine learning-based classification to detect both traditional and Zero-Font phishing attempts. At its core, the system incorporates an advanced machine learning model trained on a well-established dataset comprising both phishing and legitimate emails. The model alone achieves an accuracy rate of up to 98.8%, contributing significantly to the overall effectiveness of the tool. This hybrid approach enhances the system’s robustness and detection accuracy across diverse phishing scenarios. The findings underscore the importance of multi-faceted detection mechanisms and contribute to the development of more resilient defenses in the ever-evolving landscape of cybersecurity threats. Full article

Privacy-preserving record linkage (PPRL) aims to link records from different data sources while ensuring sensitive information is not disclosed. Utilizing blockchain as a trusted third party is an effective strategy for enhancing transparency and auditability in PPRL. However, to ensure data privacy during computation, such approaches often require computationally intensive cryptographic techniques. This can introduce significant computational overhead, limiting the method’s efficiency and scalability. To address this performance bottleneck, we combine blockchain with the distributed computation of secret sharing to propose a PPRL method based on blockchain-coordinated distributed computation. At its core, the approach utilizes Bloom filters to encode data and employs Boolean and arithmetic secret sharing to decompose the data into secret shares, which are uploaded to the InterPlanetary File System (IPFS). Combined with masking and random permutation mechanisms, it enhances privacy protection. Computing nodes perform similarity calculations locally, interacting with IPFS only a limited number of times, effectively reducing communication overhead. Furthermore, blockchain manages the entire computation process through smart contracts, ensuring transparency and correctness of the computation, achieving efficient and secure record linkage. Experimental results demonstrate that this method effectively safeguards data privacy while exhibiting high linkage quality and scalability. Full article

This study investigates the convergence of blockchain, artificial intelligence (AI), near-field communication (NFC), and mobile technologies in electronic payment (e-payment) systems, proposing an innovative integrative framework to deconstruct the systemic innovations and transformative impacts driven by such technological synergy. Unlike prior research, which often focuses on single-technology adoption, this study uniquely adopts a cross-technology convergence perspective. To our knowledge, this is the first study to empirically map the multi-technology convergence landscape in e-payment using scientometric techniques. By employing bibliometric and thematic network analysis methods, the research maps the intellectual evolution and key research themes of technology convergence in e-payment systems. Findings reveal that while the integration of these technologies holds significant promise, improving transparency, scalability, and responsiveness, it also presents challenges, including interoperability barriers, privacy concerns, and regulatory complexity. Furthermore, this study highlights the potential for convergent technologies to unintentionally deepen the digital divide if not inclusively designed. The novelty of this study is threefold: (1) theoretical contribution—this study expands existing frameworks of technology adoption and digital governance by introducing an integrated perspective on cross-technology adoption and regulatory responsiveness; (2) practical relevance—it offers actionable, stakeholder-specific recommendations for policymakers, financial institutions, developers, and end-users; (3) methodological innovation—it leverages scientometric and topic modeling techniques to capture the macro-level trajectory of technology convergence, complementing traditional qualitative insights. In conclusion, this study advances the theoretical foundations of digital finance and provides forward-looking policy and managerial implications, paving the way for a more secure, inclusive, and innovation-driven digital payment ecosystem. Full article

This paper presents a novel deep learning-based automatic modulation recognition (AMR) model, designed to classify ten modulation types from complex I/Q signal data. The proposed architecture, named CV-CNN-TCN, integrates Complex-Valued Convolutional Neural Networks (CV-CNNs) with Temporal Convolutional Networks (TCNs) to jointly extract spatial and temporal features while preserving the inherent phase information of the signal. An enhanced variant, CV-CNN-TCN-DCC, incorporates dilated causal convolutions to further strengthen temporal representation. The models are trained and evaluated on the benchmark RadioML2016.10b dataset. At SNR = −10 dB, the CV-CNN-TCN achieves a classification accuracy of 37%, while the CV-CNN-TCN-DCC improves to 40%. In comparison, ResNet reaches 33%, and other models such as CLDNN (convolutional LSTM dense neural network) and SCRNN (Sequential Convolutional Recurrent Neural Network) remain below 30%. At 0 dB SNR, the CV-CNN-TCN-DCC achieves a Jaccard index of 0.58 and an MCC of 0.67, outperforming ResNet (0.55, 0.64) and CNN (0.53, 0.61). Furthermore, the CV-CNN-TCN-DCC achieves 75% accuracy at SNR = 10 dB and maintains over 90% classification accuracy for SNRs above 2 dB. These results demonstrate that the proposed architectures, particularly with dilated causal convolutional enhancements, significantly improve robustness and generalization under low-SNR conditions, outperforming state-of-the-art models in both accuracy and reliability. Full article

In modern manufacturing, ensuring quality control for geometric features is critical, yet detecting anomalies in circular components remains underexplored. This study proposes a real-time defect detection framework for metal parts with holes, optimized for deployment on a Raspberry Pi 5 edge device. We fine-tuned and evaluated three deep learning models ResNet50, EfficientNet-B3, and MobileNetV3-Large on a grayscale image dataset (43,482 samples) containing various hole defects and imbalances. Through extensive data augmentation and class-weighting, the models achieved near-perfect binary classification of defective vs. non-defective parts. Notably, ResNet50 attained 99.98% accuracy (precision 0.9994, recall 1.0000), correctly identifying all defects with only one false alarm. MobileNetV3-Large and EfficientNet-B3 likewise exceeded 99.9% accuracy, with slightly more false positives, but offered advantages in model size or interpretability. Gradient-weighted Class Activation Mapping (Grad-CAM) visualizations confirmed that each network focuses on meaningful geometric features (misaligned or irregular holes) when predicting defects, enhancing explainability. These results demonstrate that lightweight CNNs can reliably detect geometric deviations (e.g., mispositioned or missing holes) in real time. The proposed system significantly improves inline quality assurance by enabling timely, accurate, and interpretable defect detection on low-cost hardware, paving the way for smarter manufacturing inspection. Full article

In the rapid process of urbanization, urban agglomerations have become a key driving factor for regional development and spatial reorganization. The formation and development of urban agglomerations rely on communication between cities. However, the spatiotemporal characteristics of intercity travelers are not fully grasped throughout the entire trip chain. This study proposes a spatiotemporal analysis method for intercity travel in urban agglomerations by constructing origin-to-destination (OD) trip chains using smartphone data, with the Beijing–Tianjin–Hebei urban agglomeration as a case study. The study employed Cramer’s V and Spearman correlation coefficients for multivariate feature selection, identifying 12 key variables from an initial set of 20. Then, optimal cluster configuration was determined via silhouette analysis. Finally, the K-prototypes algorithm was applied to cluster 161,797 intercity trip chains across six transportation corridors in 2019 and 2021, facilitating a comparative spatiotemporal analysis of travel patterns. Results show the following: (1) Intercity travelers are predominantly males aged 19–35, with significantly higher weekday volumes; (2) Modal split exhibits significant spatial heterogeneity—the metro predominates in Beijing while road transport prevails elsewhere; (3) Departure hubs’ waiting times increased significantly in 2021 relative to 2019 baselines; (4) Increased metro mileage correlates positively with extended intra-city travel distances. The results substantially contribute to transportation planning, particularly in optimizing multimodal hub operations and infrastructure investment allocation. Full article

This research evaluates the performance of six metaheuristic algorithms in the active and reactive power management of wind turbines (WTs) integrated into an AC microgrid (MG). The population-based genetic algorithm (PGA) is proposed as the primary optimization strategy and is rigorously compared against five benchmark techniques: Monte Carlo (MC), particle swarm optimization (PSO), the JAYA algorithm, the generalized normal distribution optimizer (GNDO), and the multiverse optimizer (MVO). This study aims to minimize, through independent optimization scenarios, the operating costs, power losses, or CO₂ emissions of the microgrid during both grid-connected and islanded modes. To achieve this, a coordinated control strategy for distributed generators is proposed, offering flexible adaptation to economic, technical, or environmental priorities while accounting for the variability of power generation and demand. The proposed optimization model includes active and reactive power constraints for both conventional generators and WTs, along with technical and regulatory limits imposed on the MG, such as current thresholds and nodal voltage boundaries. To validate the proposed strategy, two scenarios are considered: one involving 33 nodes and another one featuring 69. These configurations allow evaluation of the aforementioned optimization strategies under different energy conditions while incorporating the power generation and demand variability corresponding to a specific region of Colombia. The analysis covers two-time horizons (a representative day of operation and a full week) in order to capture both short-term and weekly fluctuations. The variability is modeled via an artificial neural network to forecast renewable generation and demand. Each optimization method undergoes a statistical evaluation based on multiple independent executions, allowing for a comprehensive assessment of its effectiveness in terms of solution quality, average performance, repeatability, and computation time. The proposed methodology exhibits the best performance for the three objectives, with excellent repeatability and computational efficiency across varying microgrid sizes and energy behavior scenarios. Full article

Microservice architecture (MSA) has garnered attention in various software communities because of its significant advantages. Organizations have also prioritized migrating their legacy systems to MSA, seeking to gather the intrinsic advantages of this architectural style. Despite the importance of this architectural style, there is a lack of comprehensive studies in the literature on the modernization of legacy systems to MSA. Thus, the principal objective of this article is to present a comprehensive overview of this research theme through a mixed-method investigation composed of a systematic mapping study based on 43 studies and an empirical evaluation by industry practitioners. From these, a taxonomy for the initiatives identified in the literature is established, along with the application domain for which such initiatives were designed, the methods used to evaluate these initiatives, the main quality attributes identified in our investigation, and the main activities employed in the design of such initiatives. As a result, this article delineates a process of modernization based on six macro-activities, designed to facilitate the transition from legacy systems to microservice-based ones. Finally, this article presents a discussion of the results based on the evidence gathered during our investigation, which may serve as a source of inspiration for the design of new initiatives to support software modernization. Full article