Conference Agenda

Liquefied petroleum gas transportation in tank trucks represents a critical risk due to leaks that can cause explosions, such as the Villa El Salvador accident (2020) with 34 fatalities. This study developed and experimentally validated an early detection system based on the MQ-5 sensor and Arduino Uno microcontroller. A total of 120 trials were conducted at four distances (1, 5, 10, 15 cm) with 30 repetitions each, establishing a threshold of 400 ADC units. Results showed response times between 0.98 and 3.37 seconds, lower than the 10-second standard. The system achieved 92.50% accuracy with perfect 100% precision (zero false alarms) and 91% sensitivity. Analysis of variance confirmed that distance significantly affects response time (F=557.03, p<0.000001), establishing a predictive model (R²=0.9985): Time = 0.171×Distance + 0.629. Energy consumption (<250 mW) and low cost (~$40 USD) confirm its viability for vehicular applications, providing the first scientific quantification of the distance-time relationship for MQ-5 sensors.

This study aimed to implement the Power BI business intelligence tool to improve purchasing management in the logistics department of a sulfate production company. The research focuses on the logistics purchasing process to measure how the use of a business intelligence solution impacts procurement, information quality, and product control and tracking. Methodologically, the study employed a quantitative, applied, and quasi-experimental design with a population of 19 logistics employees. Data was collected using a structured questionnaire with 15 items. The results obtained from the statistical analysis showed a 75% improvement in the purchasing process, as well as a 30% reduction in processing time, thus optimizing the company's logistics management.

This research aimed to analyze, through a systematic literature review, the applications of Artificial Intelligence (AI) geared towards reducing the carbon footprint of computing systems. To this end, a methodology based on the PRISMA guidelines was applied, conducting searches in high- impact academic databases such as Scopus, considering articles published between 2021 and 2025 related to AI, sustainability, energy efficiency, green software, and data centers. After the identification, screening, and eligibility process, 31 articles were selected that met the inclusion criteria defined by the PICO model. The results showed that AI is applied in multiple areas with significant improvements in energy efficiency and emissions reduction. In data centers, it is used to optimize energy consumption and thermal management, resource allocation, and real-time electricity consumption. In sustainable software, optimization algorithms and machine learning models are applied to reduce computational demand. In sectors such as industry, agriculture, and logistics, predictive models and hybrid techniques show significant reductions in CO₂ emissions and improvements in operational efficiency. Likewise, the growing adoption of digital twins, deep learning, and approaches based on dynamic optimization was identified. Despite this progress, significant gaps remain, as most research focuses on isolated sectoral applications, and a consolidated analysis of AI's contribution to computational systems specifically aimed at reducing carbon footprints is still lacking. In conclusion, these findings support the need for new strategic guidelines that incorporate dispersed contributions, synthesize trends, and establish clear and concise guidelines for developing more efficient computational architectures.

Changing the way crime is investigated in the mobile era requires rigor and humanity: recovering the stories stored in phones without compromising their integrity. With that premise, this review sought to identify how effective and reliable current mobile forensic methodologies and tools are, and how they are being used in real investigative contexts. A structured systematic review was conducted using the PICO model, a Scopus search, and a PRISMA flowchart: from 229 records, 20 recent studies were screened and evaluated, examining research designs, methodological frameworks, extraction techniques, and performance by application and operating system. The results reveal an operational consensus: DFRWS and NIST frameworks guide the sequential phases of identification, preservation, acquisition, analysis, and reporting; commercial forensic suites (Cellebrite UFED, Magnet AXIOM, XRY, Oxygen, Belkasoft, MOBILedit) achieve high recovery rates, although with heterogeneous performance depending on the app, version, and encryption; and emerging automation approaches (ML/LLM and multimodal semantic analysis) accelerate triage and improve the correlation of textual, audio, image, and video evidence. Limitations persist, including modern encryption, cloud dependencies, uneven parsers, and explainability issues that affect legal admissibility. In conclusion, the evidence supports an integrated model—“standard framework + validated tool + explainable automation”—sustained by strict chain of custody and reproducible documentation. The review recommends open test datasets, cross-tool validation by case/OS/encryption level, and governance/ethical guidelines that balance investigative efficiency with privacy protection.

The growing integration of artificial intelligence (AI)-based applications into everyday contexts has extended their use to the family environment, where technological adoption is often mediated by responsible adults. The present study analyzed parental perceptions of the use of AI-based applications, incorporating contextual information about the usage scenarios reported by parents. A cross-sectional observational design was used, employing a self-administered online questionnaire applied to 303 parents residing in Honduras. Parental perception was assessed using a six-item Likert scale, while the age of the children and the type of AI-based application were used for descriptive and contextual purposes. Analyses included descriptive statistics, Welch's analysis of variance, and latent profile analysis (LPA). The results showed a moderate overall perception of the impact of using AI applications, with no significant differences based on the gender of the participants, but with relevant variations based on the age group of the parents. The LPA identified three distinct profiles of parental perception: low, moderate, and high perceived impact, suggesting heterogeneous configurations of technological acceptance. These findings highlight the role of parents as key mediators in the adoption of intelligent systems and underscore the importance of considering human and perceptual factors in the design of AI-based applications. The implications of the study are relevant for the development of user-centered technological solutions that are sensitive to family contexts.

Unplanned diesel-engine failures lead to relevant operational and economic impacts in heavy-transport settings. Predictive maintenance approaches based on remaining useful life (RUL) estimation can mitigate downtime; however, real labeled degradation datasets are often scarce. This work presents a simulation-based, reproducible framework to generate multivariate synthetic OBD-II–like signals and evaluate RUL prediction using a Long Short-Term Memory (LSTM) network. In addition, a Genetic Algorithm (GA) is applied to search LSTM hyperparameters using validation performance as the fitness criterion. The system is assessed as a regression problem using 𝑹𝟐, RMSE, MAE, and relative RMSE under an asset-based split, where an unseen engine is reserved for testing. Results show that both the baseline LSTM and the GA-tuned LSTM satisfy the predefined acceptance criteria (𝑹𝟐 ≥ 𝟎. 𝟖𝟓 and relative RMSE ≤ 𝟏𝟓%), while the baseline model achieves a lower test RMSE than the GA-tuned configuration, highlighting that improved validation fitness does not necessarily translate into better generalization on unseen assets. The proposed framework provides a controlled proof of concept for future studies incorporating real OBDII data and broader validation.

This study developed a convolutional neural network for the automated detection of dental conditions—specifically caries, restorations, root canals, and prostheses—on radiographic images using the Roboflow platform and deployed on the NVIDIA Jetson Nano. A dataset of 1,933 dental X-ray images, provided by the Bright Smile clinic and expanded to 4,639 through data augmentation techniques, was used. The model was trained and tested in five iterative versions, gradually incorporating each dental condition class. Preprocessing techniques, including a 20% increase in image saturation and targeted augmentations, significantly improved detection performance. The final model achieved a mean Average Precision (mAP) of 97.7%, with notable improvement in the identification of dental caries—previously the most challenging class. These results demonstrate that optimized preprocessing combined with YOLO-based training in Roboflow and deployment on Jetson Nano constitutes an effective pipeline for real-time dental diagnostics.

This article presents the development of a preventive maintenance plan for the escalators of the Panama Metro. It addresses the technical and functional specifications of the electromechanical escalators on Line 1 of the Panama Metro and the corresponding technical standards ISO 25745 and EN 115. A literature review of other maintenance plans for escalators on the metro line is conducted, and a proposal for a preventive maintenance plan and its scheduling for the electromechanical escalators of the Panama Metro is developed.