Eduardo Avendaño Fernández

Context: This investigation assessed whether repeated exposure to visual-spatial patterns and guided micro-interactions could lead high school and first-year engineering students to internalize formal rules. By combining implicit learning tools (e.g., VSRT) with explicit tools, such as a binary tree calculator. Objective: This study aims to determine whether stable patterns and low-friction actions increase implicit learning, structured reasoning, and algebraic manipulation compared with historical baseline performance. Methodology: A concurrent mixed-methods design was used to recruit 40 participants from a purposive historical sample of 153 students. Quantitative (VSRT accuracy, interaction statistics) and qualitative (28 interviews until theoretical saturation) analyses were performed using thematic analysis and decision trees. Results: The participants achieved an accuracy of 84% in recognizing VSRT patterns, exceeding the previous average of 70% and showing clear improvement. Repetition and familiarity with tasks were found to be important factors that influenced progress. Conclusions: Implicit VSRT training combined with explicit visualization is an efficient method for teaching hierarchical content. Such dual approaches enhance academic performance as well as structural understanding and can be evidenced by repetition and familiarity as pivotal techniques. This study provides a strong model for STEM education interventions that link implicit and explicit cognitive systems.

Reliable person segmentation with sparse 3D sensors degrades significantly under adverse atmospheric conditions. This work presents a controlled comparative evaluation of four segmentation architectures—U-Net, Mask R-CNN, DeepLabV3+, and FCN-ResNet—on a fused Radar–LiDAR dataset for binary person–background segmentation and applies a dual-domain evaluation procedure that formally links 2D pixel-level overlap (IoU, Dice) to 3D geometric fidelity (Chamfer distance, Completeness) through mask back-projection onto fused point clouds. Raw point clouds are rasterized into range–intensity grids enriched with Radar reflectivity; the predicted masks are then reprojected into 3D space and evaluated using Chamfer distance and Completeness under three controlled visibility conditions. U-Net achieves the highest 2D overlap (IoU = 0.82, Dice = 0.89), while DeepLabV3+ delivers the best 3D reconstruction fidelity (Chamfer = 0.021 m, Completeness = 93.4%) and the highest overall accuracy (97.9%). This dissociation between 2D overlap and 3D fidelity is explained by DeepLabV3+’s multi-scale Atrous Spatial Pyramid Pooling (ASPP), which reduces boundary fragmentation during back-projection; more than 70% of the Chamfer deviation across competing architectures originates at object contours. Mask R-CNN performs well when instances are clearly separated, and FCN-ResNet offers the lowest computational cost at reduced boundary precision. Radar–LiDAR fusion sustains an IoU within 3% of clear-weather performance under dense fog, whereas LiDAR-only inputs degrade by more than 12%. Due to the 12:1 background-to-person class imbalance, overlap-based metrics (IoU, Dice) are prioritized over raw accuracy in all reported comparisons. These results provide actionable deployment guidance and constitute a reproducible evaluation procedure for future sparse-sensor fusion studies, independently of the architectures evaluated.

Wireless sensor networks (WSN) are essential for monitoring operations, environmental conditions, and ensuring worker safety in underground mining environments. A key challenge in deploying these networks lies in achieving sufficient sensor coverage and reliable connectivity to enable early incident detection and timely response. While existing approaches primarily focus on deployment efficiency, this paper introduces a safety-driven optimization model that accounts for the complete underground mine structure and its specific operational conditions. The proposed algorithm minimizes the number of deployed nodes—reducing capital expenditure (CAPEX)—while meeting occupational safety standards by ensuring sensor coverage at critical points of interest and maintaining network connectivity. Compared to existing models, our method reduces the number of required nodes by up to 82%, while satisfying both k-coverage and k-connectivity requirements. Unlike conventional strategies based on idealized assumptions, our approach is tailored to the real-world conditions of Colombian underground mines. Simulation results demonstrate that achieving connectivity demands more nodes than coverage when both are evaluated under equivalent conditions. The algorithm was validated in two prototype mines, achieving k = 2 for both coverage and connectivity—ensuring that each point of interest is monitored by at least two sensors and connected via two link-disjoint communication paths. This redundancy enhances fault tolerance by providing backup for both sensing and communication. Finally, we introduce two new metrics for evaluating k-coverage and k-connectivity, designed to assess the robustness of WSN in underground mining and tunnel environments.

Introduction: In engineering courses being able to connect theory with practice to validate concepts and solve problems from acquired knowledge, critical thinking and creativity, is a challenge. In project-based learning and active learning, it has been validated that developing programming activities in real applications allows appropriate concepts, developing skills and soft competences. Objective: to diagnose the current state of the learning outcomes in the core course of an electronic engineering curriculum in order to propose models to achieve the learning objectives. Methodology: a diagnostic instrument was designed for undergraduate students in electronic engineering, building a matrix of Weaknesses, Opportunities, Strengths and Opportunities (SWOT) to apply data analytics and generate diagrams of co-occurrence/frequency of discourse and from which the results are discussed. The instrument was applied to a cumulative sample of 87 students, identifying keywords related to the use of tools, previous knowledge (mathematical foundation), and experimental work at laboratory. Results: The analysis identified weaknesses in some concepts, deficiencies to operationalize, and the need to complement experimentally, allowing adjusting curricular contents and demonstrating the application of learning models and methodologies. Conclusion: Interdisciplinary topics enable the design of challenging projects to assess competencies and learning outcomes, validating an active and project-based learning approach (A+PBL) in a pilot Digital Signal Processing (DSP) course.

The use of LiDAR technologies offer an effective alternative for enhancing monitoring and surveillance tasks in secured areas, or hard-to-access areas. Conventional solutions allow detection and identification capabilities, wich are valuable for object and pattern recognition in environments with special conditions. This paper proposes a method for structure a dataset by collecting data from LiDAR sensors and reconstructing 3D point cloud scenes into images. Furthermore, training with these data using semantic segmentation algorithms allows for the detecting of people in monitoring and surveillance contexts. The process involves data acquisition techniques, preprocessing, and the application of tools and technologies for semantic segmentation in the post-processing stage. Data is acquired from RPLidar sensor with a spatial resolution of less than 0.4 degrees per step, ensuring greater accuracy in detecting small objects. Scanning distances between 0.15 and 12 meters allow for the capture of 941 points per rotation angle. The dataset is complemented with labels and images containing binary masks of regions of interest for people detection using semantic segmentation. The main contribution of this paper is the methodological description of constructing a dataset comprising 1,254 images in various scenarios, including indoors and outdoors, with different levels of occlusion and illumination. This dataset facilitates the development of machine learning algorithms and deep learning algorithms for pattern recognition, such as Mask R-CNN (89.9% mean class accuracy) and U-NET (90.5% mean class accuracy). The results demonstrated an approximate 2% improvement in mean class accuracy compared to other state-of-the-art algorithms in semantic segmentation for point clouds.

This chapter introduces the foundations of the Extended Kalman Filter (EKF) and Unscented Kalman Filter (UKF) for data sensor fusion applications in surveillance applications. After explaining how to model the drone under the constant turn rate and velocity (CTRV) dynamics, the EKF and UKF techniques for Lidar and Radar data sensor fusion are applied to enable better 3D object detection and reconstruction from point cloud data is evaluated under a performance comparison. Through root mean square error (RMSE) and mean square error (MSE) the filters are designed to meet the specifications of the model based on the position of the gathered points (X and Y axis) in the space state model. By degrading the measurements with Gaussian and Rayleigh noise, the results evaluate and compare both techniques to show the tolerance and performance in terms of stability, scalability, and application to drone monitoring in surveillance activities.

This paper presents a novel methodology for deploying wireless sensor nodes in the Industrial Internet of Things (IIoT) to address the safety and efficiency challenges in underground coal mining. The methodology is intended to support long-term planning on mitigating the risks in occupational health and safety policies. To ensure realistic and accurate deployment, we propose a software tool that generates mine models based on geolocation data or blueprints in image format, allowing precise adaptation to the specific conditions of each mine. Furthermore, the process is based on sensing and communication range values obtained through simulations and on-site experiments. The deployment strategy is articulated in two complementary steps: a deterministic deployment, where nodes are strategically placed according to the structure of the tunnels, followed by a random stage to include additional nodes that ensure optimal coverage and connectivity inside the mine by comparing different methodologies for deploying sensor networks using coverage density as a performance metric. We analyze coverage and connectivity based on the three probability density functions (PDFs) for the random deployment of nodes: uniform, normal, and exponential, evaluating both the degree of coverage (k-coverage) and the degree of connectivity (k-connectivity). The results show that our proposed methodology stands out for its lower density of sensors per square meter, which translates into a reduction of between 20.81% and 23.46% for uniform and exponential PDFs, respectively, concerning the number of sensors compared to the analyzed methodologies. In this way, it is possible to determine which distribution is suitable to cover the elongated area with the smallest number of nodes, considering the coverage and connectivity requirements, to reduce the deployment cost. The uniform PDF minimizes the number of sensors needed by 44.70% in small mines and 46.27% in medium ones compared to the exponential PDF. These findings provide valuable information to optimize node deployment regarding cost and efficiency; a uniform function is a good option depending on prices. The exponential distribution reached the highest values of k-coverage and k-connectivity for small and medium-sized mines; in addition, it has greater robustness and tolerance to faults like signal network intermittence. This methodology not only improves the collection of critical information for the mining operation but also plays a vital role in reducing the risks to the health and safety of workers by providing a more robust and adaptive monitoring system. The approach can be used to plan IIoT systems based on Wireless Sensor Networks (WSN) for underground mining exploitation, offering a more reliable and adaptable strategy for monitoring and managing complex work environments.

Este artículo presenta una revisión de estado de arte de las técnicas del aprendizaje de máquina para el reconocimiento de patrones, la definición de los algoritmos para la clasificación de objetos (personas) a partir de imágenes construidas desde nubes de puntos adquiridas por sensores de Radar y Lidar, y resultados preliminares de aplicación en escenas reconstruidas a partir de nubes de puntos. En el primer hito del proyecto CLARIFIER (frequenCy-agiLe rAdar-lidaR chIp For surveIllancE moving platfoRms), se ha estudiado e implementado un filtro de Kalman Extendido para fusión de datos de sensores de Radar y LiDAR, cuyo aporte fue la inclusión de la velocidad angular en el modelo cinemático de un drone. En el segundo hito, se ha revisado estado de arte para identificar enfoque (segmentación semántica) y algoritmos que permitan detectar personas en imágenes reconstruidas de escenarios o zonas con requerimientos de supervisión y monitoreo. Como resultado de aplicación preliminar, se ha construido un conjunto de datos de escenas que incluye la clase “personas” y se evaluaron los algoritmos U-Net y Mask R-CNN que aplicas técnicas de segmentación semántica. Dado que la resolución del sensor en particular de Lidar, es mayor a la del radar, así como la diferencia en rangos que cada uno alcanza, se genera un compromiso que requiere múltiples barridos de la escena desde diferentes ángulos y distancias y se requiere aumentar el conjunto de datos, para entrenar los algoritmos y mejorar el porcentaje de detecciones correctas. De acuerdo a las métricas que se han evaluado (pérdidas, exactitud, precisión, exhaustividad y F1 Score), la clasificación de la clase personas en el contexto de segmentación semántica alcanza un 89.91 % para Mask R-CNN, y del 90.53% para U-Net; y con la curva de operación característica del receptor (ROC) y el área bajo la curva (AUC) se obtiene un 90% y 92% en la detección de la clase personas, respectivamente. Este resultado valida la efectividad de estos modelos de redes neuronales convolucionales aplicada a imágenes obtenidas a partir de nubes de puntos.

In surveillance and monitoring systems, the use of mobile vehicles or unmanned aerial vehicles (UAVs), like the drone type, provides advantages in terms of access to the environment with enhanced range, maneuverability, and safety due to the ability to move omnidirectionally to explore, identify, and perform some security tasks. These activities must be performed autonomously by capturing data from the environment; usually, the data present errors and uncertainties that impact the recognition and resolution in the detection and identification of objects. The resolution in the acquisition of data can be improved by integrating data sensor fusion systems to measure the same physical phenomenon from two or more sensors by retrieving information simultaneously. This paper uses the constant turn and rate velocity (CTRV) kinematic model of a drone but includes the angular velocity not considered in previous works as a complementary alternative in Lidar and Radar data sensor fusion retrieved using UAVs and applying the extended Kalman filter (EKF) for the detection of moving targets. The performance of the EKF is evaluated by using a dataset that jointly includes position data captured from a LiDAR and a Radar sensor for an object in movement following a trajectory with sudden changes. Additive white Gaussian noise is then introduced into the data to degrade the data. Then, the root mean square error (RMSE) versus the increase in noise power is evaluated, and the results show an improvement of 0.4 for object detection over other conventional kinematic models that do not consider significant trajectory changes.

Se presenta un instrumento para evaluar competencias pedagógicas, didácticas y en tecnología de la información y las comunicaciones, en el marco de la cuarta revolución industrial y definida dentro de las principales tendencias de innovación y cambio en la era del conocimiento en lo que se conoce como Educación 4.0. Existen brechas importantes entre las instituciones de educación básica primaria y secundaria en los sectores público y privado, rural y urbano, y municipios certificados y no certificados. El resultado de las pruebas Pisa a nivel Colombia, identifica unas debilidades en la educación, que fue un reto en la adopción de estrategias conectadas o desconectadas al proceso de enseñanza por la condición de la pandemia COVID-19, y por la diferencia en la formación, conceptualización y aplicación de los objetivos y resultados de aprendizaje desde el profesorado. Aunque de manera ágil se generó un proceso de adaptación y ajuste en la metodología de enseñanza (acceso remoto síncrono), también se identificó una serie de oportunidades para fortalecer los procesos de enseñanza-aprendizaje, desde la docencia e infraestructura, y las capacidades de laboratorio para pasar a un modelo de aprendizaje centrado en el estudiante. Se debe promover las competencias blandas para ofrecer solución a los problemas, desarrollar el pensamiento crítico, incentivar el trabajo interdisciplinar y cooperativo, y la aplicación de nuevos modelos pedagógicos como el aprendizaje basado en proyectos, en problemas, experiencias, etc. Se presenta la hipótesis de que promover el pensamiento computacional en una ruta hacia la educación STEAM+H que permita un diseño curricular para un aprendizaje que ofrezca soluciones transformadoras, innovadoras y creativas en contexto y que atienda los retos globales en los objetivos de desarrollo sostenible. El diseño del instrumento que se propone, pretende ir más allá de la caracterización del profesorado, cuya población objetivo será las instituciones de educación básica y secundaria (posteriormente podrá ser adaptado para educación superior). El instrumento se desglosa por subáreas dentro de la pedagogía, la didáctica y las TIC, desde la realidad del territorio y ajustado a las brechas de acceso a la información y limitantes socio-económicas de las instituciones. Se considera que el aporte al conocimiento se consolida desde un diagnóstico, para potenciar las capacidades docentes como actores cruciales en el proceso disruptivo en esta nueva revolución educativa que demanda habilidades y aplicación de conocimiento. En esta ruta de aprendizaje y enseñanza, la formación a lo largo del proceso educativo permitirá apropiar las herramientas necesarias por los estudiantes para que puedan usarlas y con la fortaleza de que los futuros aspirantes a la Universidad cuenten con una mayor capacidad de análisis y de solución de problemas, siendo éste un reto para el modelo educativo del país por las brechas identificadas en el contexto nacional. El instrumento está en etapa de validación previo a un piloto en Boyacá a nivel de educación pública como preámbulo en el programa de territorio STEAM, con el apoyo de programas de la UPTC, en la construcción de una política educativa para el territoriol.