Abstracts:This paper introduces the application of novel eyetracking metrics to assess visual attention and cognitive load in neonatal pain assessment. Our goal is to evaluate pediatrician experts, non-experts, and parents using the relative Explore- Exploit Ratio, along with the Task-Evoked Pupillary Response, while analyzing the frontal faces of distinct newborns before and after painful procedures. All the experiments were based on a benchmark image dataset considering clinically relevant areas of interest. The Tobii TX300 system was used to record the eye-tracking data in a closed room with controlled lighting. Our results disclose that the visual attention described by the traditional metrics does not correspond directly to the respective fixation patterns and pupillary changes quantified for all the sample groups of participants investigated, highlighting statistically significant differences in the visual behavior between participants with or without clinical experience only when using the novel metrics proposed instead.

Abstracts:FrodoKEM, a key encapsulation mechanism (KEM) based on the learning with errors (LWE) problem, would be included for standardization by the International Organization for Standardization (ISO) and recommended for PQC migration by the BSI (German Federal Office for Information Security) and the ANSSI (French Cybersecurity Agency). It is closely related to the challenging time-computational problem inherent to algebraically unstructured lattices. However, hardware implementations of this scheme are required to verify its effectiveness in real-world applications. To the best of our knowledge, this is the first hardware implementation of FrodoKEM using High-Level Synthesis (HLS), which meets all requirements of the version submitted for standardization to ISO. The proposed design started with the profiling of the reference C software implementation using Valgrind software tools, to identify the functions that are the most time-consuming. The advantages of the proposed implementation include a 34% improvement in the speed metric of the Key Generation module in comparison with the reference software implementation. The results show that the key generation, encapsulation, and decapsulation use 26%, 39%, and 32%, respectively, of the total area utilization on the Artix-7.

Abstracts:In evaluating and validating a physical system, real-time Hardware in the Loop (HIL) emulation offers advantages such as time and cost reduction, fault prevention, and the ability to conduct validations in an environment similar to its final application. On the other hand, low-cost technologies such as microcontrollers, digital signal processors, and FPGAs have been employed to leverage the advantages of HIL emulation in the teaching process. This article describes the implementation of a low-cost, open-access HIL didactic platform for use in subjects such as differential equations, systems dynamics, and control systems, among others. The platform is based on a Raspberry Pi Pico development board and features a graphical user interface (GUI). In the GUI, the user can visualize graphs of the emulated systems variables and real-time animation of its state and export the acquired data to a comma-separated file. The functionalities offered by the platform make it an affordable tool that allows users to evaluate the response of a dynamic system, whether it is open-loop or closed-loop, without the need for classrooms or specialized equipment. Unlike similar works where HIL techniques with low-cost hardware are employed for educational purposes, the proposal in this work is more cost-effective and integrates the described GUI. All the files necessary for implementing the didactic platform are openly available in a public repository, including those needed for PCB manufacturing.

Abstracts:Photovoltaic systems must be highly efficient to transfer the electric power generated to the local loads or to the electrical grid. In partial-shading conditions there are multiple local maximum power points which are not evaluated by conventional MPPT methods producing low efficiency in the photovoltaic system. In this paper a hybrid Particle Swarm Optimization (PSO) based method is proposed. The proposed method allows to locate the global maximum power point enhancing the availability of the generated electrical power and reducing the convergence time compared to the conventional PSO algorithm. The proposed method is compared with classical MPPT algorithms like Hill Climbing (HC), perturb and observe (P and O), incremental conductance (IncCond) and the conventional particle swarm optimization (PSO) method. The comparison is performed by means of numerical simulations and implementing an experimental platform with real photovoltaic panels.

Abstracts:The increasing demand for efficient and sustainable electric vehicle (EV) charging solutions has driven the development of advanced multi-input charger systems. This paper introduces a closed-loop EV charger powered by dual energy inputs, i.e., wireless power transfer (WPT) and photovoltaic (PV) sources. The dual-input charger integrates these energy sources to ensure stable and efficient constant voltage battery charging. The output voltage is continuously compared with the required EV battery charging voltage and regulated using a proportional-integral (PI) controller. To address the nonlinear and dynamic characteristics of the input sources, a novel Hybrid GEO-PO optimization algorithm, which combines the strengths of the Golden Eagle Optimizer (GEO) and the Puma Optimizer (PO), is proposed to determine the optimal PI controller parameters. MATLAB/Simulink simulations and experimental validation demonstrate that the Hybrid GEO-PO algorithm outperforms its parent algorithms in regulating EV battery charging voltage. The hybrid algorithm achieves faster response times, lower overshoot, and enhanced robustness compared to the standalone GEO and PO algorithms. Additionally, the successful implementation of the system using an FPGA controller highlights its practicality and suitability for real-world applications. This study establishes the Hybrid GEO-PO algorithm as a superior and promising approach for optimizing dual-input EV chargers, paving the way for next-generation intelligent charging infrastructure.

Abstracts:Forecasting electricity generation from renewable resources is crucial for the efficient planning and operation of power systems. The development of forecasting models based on local meteorological variables is common, however, sometimes this information is unavailable. This study explores the use of multivariate models that do not incorporate meteorological variables, but use historical power-generated data from eight PV plants located in the same region to predict the future value of a target plant. This allows for improved forecasting when meteorological variables are unavailable and the only information available is the generation of the PV plants. The performance of LSTM and BiLSTM networks is compared for different time horizons, considering various lags of the power series itself for estimating future values. The main contributions of this study include the introduction of power time series from other plants as model inputs, the use of spatial interpolation to fill in missing data and the application of causality tests between time series for the selection of predictor variables, and the uncertainty associated with the predictions is analyzed using quantile regression techniques.

Abstracts:Industry 4.0 has progressed rapidly due to advances in the Internet of Things (IoT), which have significantly reduced implementation costs while enhancing technological capabilities. These developments have fostered the emergence of the Industrial Internet of Things (IIoT), particularly within scientific and industrial domains. In response to the growing demand for skilled professionals in this field, this paper presents the design and implementation of an IIoT-based monitoring system for the CASIRI radio astronomy station, a mobile platform used to characterize potential sites for radio observatories in Colombia and Antarctica. The proposed system is integrated into a broader IIoT platform designed to support both operational monitoring and educational activities. The IIoT solution features real-time data acquisition, robust hardware and software components, and reliable communication protocols within a modular and scalable architecture. Its performance and component interoperability were validated through a series of experimental tests under simulated conditions. Additionally, this work introduces educational guides to facilitate hands-on training in Fourth Industrial Revolution technologies.

Abstracts:In the past few years, path planning and scheduling became a high-impact research topic due to their real-world applications such as transportation, manufacturing and robotics. This paper focuses on the Multi-robot Path Planning (MPP) problem, which consists of planning the route for a set of robots in a given static environment. The main goal is to navigate the robots from a starting point to a destination point without colliding with other robots or static obstacles. We propose a hybrid method -- H* -- that combines adaptive route planning based on A* and local search algorithm to optimize routes in the context of the MPP problem. The A* algorithm finds the optimal solution for the route search problem and a heuristic approach is applied to scale up to the multi-agent scenario.The overall length of determined paths and the number of robot collisions is minimized during the evaluations specific small-scale environments.Computational experiments are conducted for multi-robot scenarios and the performance of H* is compared to several path-searching algorithms including A* variations extended for the multi-agent scenario and coevolutionary algorithms.Experimental results demonstrate that H* outperforms the A* based heuristic approaches in terms of path length. H* shows similar performance as the coevolutionary method and performs better on smaller-scale maps.

Abstracts:In the context of autonomous navigation, the development of systems that enable vehicles to operate independently in controlled environments is a crucial step toward advancing autonomous technology. This work presents the design, implementation, and validation of a navigation system for autonomous vehicles using NeuroEvolution of Augmenting Topologies (NEAT). The primary objective was to create a vehicle capable of navigating a 2D map with a defined starting point and target. Virtual sensors enable the vehicle to identify navigable paths and boundaries. Distance metrics such as Euclidean, Manhattan, and Chebyshev were employed as reward systems, continuously calculating agent positions. The closer the vehicle is to the target, the higher its fitness score, forming the basis of the fitness function. A forced reinforcement acceleration method was designed and implemented to ensure progress when the vehicle's speed fell below 0.1, preventing it from becoming stalled. Validation tests were conducted to evaluate the system's performance under varying conditions. Results demonstrate that the autonomous vehicle can navigate the map effectively, improving its fitness score in each generation depending on the distance metric used. Chebyshev performed best in obstacle-free environments, while Euclidean excelled in the presence of obstacles. The forced reinforcement method significantly reduced the time required to achieve the target fitness. These findings provide valuable insights for researchers aiming to develop NEAT-based navigation systems for autonomous vehicles.

Abstracts:In recent years, videos have been susceptible not only to any edition but also to a variety of forgeries. One of the most popular video forgeries is frame deletion, in which a group of frames is removed to hide specific actions from the human eye. When frame deletion occurs, videos selected as evidence lose their evidentiary value. This highlights the necessity of automation, especially for analyzing large volumes of videos. Thus, we measure the performance of two deep learning approaches for frame deletion detection. Both of them use Convolutional Neural Networks (CNN): The first one, a supervised 3DCNN model and, the second one, is an unsupervised model compound of VGG-16 and Resnet-50. We evaluated them using 10-fold cross-validation in the following datasets: UCF-101, VIFFD and DTD (Driving Test Dataset), which is our contribution to the data community. To the best of our knowledge, no comparison of both approaches using 10-fold cross-validation has been found in the literature before. Afterward, we analyze the results and make recommendations for future work in this area.