ISEL - Eng. Elect. Tel. Comp. - Artigos
URI permanente para esta coleção:
Navegar
Entradas recentes
- Sparse mixture of experts enhanced transformer architecture for short-term hydroelectric reservoir volume predictionPublication . Seman, Laio Oriel; Yow, Kin-Choong; Stefenon, Stefano FrizzoIn hydroelectric-based systems, effective energy generation planning relies heavily on precise forecasting of reservoir water levels. This paper proposes a novel hybrid forecasting framework that integrates multiple preprocessing strategies with a sparse Mixture of Experts enhanced Transformer architecture for short-term reservoir volume prediction. When evaluated on 19 interconnected reservoirs across two major river basins in southern Brazil using real operational data from the Brazilian National System Operator, the proposed model achieves a mean squared error of 0.062 and a mean absolute error of 0.145. Comprehensive benchmarking against 18 state-of-the-art deep learning methods demonstrates that the proposed approach significantly outperforms existing methods while maintaining computational efficiency through sparse expert routing. Our results confirm that combining diverse preprocessing strategies with conditional computation mechanisms provides superior forecasting accuracy for reservoir management in hydroelectric power systems.
- Fourier-enhanced sequence-to-sequence latent graph neural networks for multi-node spatiotemporal forecasting in a hydroelectric reservoirPublication . Seman, Laio Oriel; Stefenon, Stefano Frizzo; Yow, Kin-Choong; Coelho, Leandro dos Santos; Mariani, Viviana CoccoThis paper presents a Fourier-enhanced dynamic sequence-to-sequence latent graph neural network (Seq2SeqLatentGNN), a deep learning architecture for multi-node spatiotemporal forecasting in hydroelectric reservoir systems. The model integrates three key components: (i) a custom Fourier layer that analyzes global temporal patterns through frequency-domain transformations, (ii) a latent correlation graph convolutional network that infers relational structures between monitoring stations without requiring predefined adjacency matrices, and (iii) an attention-based sequence-to-sequence model that processes temporal dependencies while enabling multi-step forecasting. The architecture simultaneously learns graph structure and forecasting tasks, adapting to changing spatial relationships between reservoir nodes. The proposed architecture was evaluated using a comprehensive dataset derived from 19 interconnected hydroelectric reservoirs located in southern Brazil. The dataset encompasses multiple years of high-resolution (hourly) measurements, including reservoir water levels, inflow and outflow rates, precipitation records, and energy production metrics. Experimental results demonstrate that Seq2SeqLatentGNN achieves superior performance compared to conventional statistical models and contemporary machine learning methods, as measured by standard error metrics. Analysis of the learned latent correlations reveals meaningful spatial dependencies that align with hydrological principles. The model exhibits consistent performance across varying temporal patterns, adapts to regime transitions, and captures both periodic and nonstationary dynamics. The proposed architecture contributes to spatiotemporal forecasting by combining spectral processing, dynamic graph learning, and sequence modeling in a unified framework applicable to systems with evolving connectivity patterns.
- Fast and accurate system for onboard target recognition on raw SAR echo dataPublication . Jacinto, Gustavo; Véstias, Mário; Flores, Paulo; Duarte, RuiSynthetic Aperture Radar (SAR) onboard satellites provides high-resolution Earth imaging independent of weather conditions. SAR data are acquired by an aircraft or satellite and sent to a ground station to be processed. However, for novel applications requiring real-time analysis and decisions, onboard processing is necessary to escape the limited downlink bandwidth and latency. One such application is real-time target recognition, which has emerged as a decisive operation in areas such as defense and surveillance. In recent years, deep learning models have improved the accuracy of target recognition algorithms. However, these are based on optical image processing and are computation and memory expensive, which requires not only processing the SAR pulse data but also optimized models and architectures for efficient deployment in onboard computers. This paper presents a fast and accurate target recognition system directly on raw SAR data using a neural network model. This network receives and processes SAR echo data for fast processing, alleviating the computationally expensive DSP image generation algorithms such as Backprojection and RangeDoppler. Thus, this allows the use of simpler and faster models, while maintaining accuracy. The system was designed, optimized, and tested on low-cost embedded devices with low size, weight, and energy requirements (Khadas VIM3 and Raspberry Pi 5). Results demonstrate that the proposed solution achieves a target classification accuracy for the MSTAR dataset close to 100% in less than 1.5 ms and 5.5 W of power.
- Gait cycle duration analysis in lower limb amputees using an IoT-based photonic wearable sensor: a preliminary proof-of-concept studyPublication . Alves, Bruna; Fantoni, Alessandro; Matos, José; Costa, João; Vieira, ManuelaThis study represents a preliminary proof of concept intended to demonstrate the feasibility of using a single-point LiDAR sensor for wearable gait analysis. The study presents a low-cost wearable sensor system that integrates a single-point LiDAR module and IoT connectivity to assess Gait Cycle Duration (GCD) and gait symmetry in real time. The device is positioned on the medial side of the calf to detect the contralateral limb crossing—used as a proxy for mid-stance—enabling the computation of GCD for both limbs and the derivation of the Symmetry Ratio and Symmetry Index. This was conducted under simulated walking at three cadences (slow, normal and fast). GCD estimated by the sensor was compared against the visual annotation with Kinovea®, showing reasonable agreement, with most cycle-wise relative differences below approximately 13% and both methods capturing similar symmetry trends. The wearable system operated reliably across different speeds, with an estimated materials cost of under 100 € and wireless data streaming to a cloud dashboard for real-time visualization. Although the validation is preliminary and limited to a single healthy participant and a video-based reference, the results support the feasibility of a photonic, IoT-based approach for portable and objective gait assessment, motivating future studies with larger and clinical cohorts and gold-standard references to quantify accuracy, repeatability and clinical utility.
- Decentralized multi-agent reinforcement learning with visible light communication for robust urban traffic signal controlPublication . Augusto Vieira, Manuel; Gonçalo Galvão; Vieira, Manuela; Véstias, Mário; Louro, Paula; Vieira, PedroThe rapid growth of urban vehicle and pedestrian flows has intensified congestion, delays, and safety concerns, underscoring the need for sustainable and intelligent traffic management in modern cities. Traditional centralized traffic signal control systems often face challenges of scalability, heterogeneity of traffic patterns, and limited real-time adaptability. To address these limitations, this study proposes a decentralized Multi-Agent Reinforcement Learning (MARL) framework for adaptive traffic signal control, where Deep Reinforcement Learning (DRL) agents are deployed at each intersection and trained on local conditions to enable real-time decision-making for both vehicles and pedestrians. A key innovation lies in the integration of Visible Light Communication (VLC), which leverages existing LED-based infrastructure in traffic lights, streetlights, and vehicles to provide high-capacity, low-latency, and energy-efficient data exchange, thereby enhancing each agent’s situational awareness while promoting infrastructure sustainability. The framework introduces a queue–request–response mechanism that dynamically adjusts signal phases, resolves conflicts between flows, and prioritizes urgent or emergency movements, ensuring equitable and safer mobility for all users. Validation through microscopic simulations in SUMO and preliminary real-world experiments demonstrates reductions in average waiting time, travel time, and queue lengths, along with improvements in pedestrian safety and energy efficiency. These results highlight the potential of MARL–VLC integration as a sustainable, resilient, and human-centered solution for next-generation urban traffic management.
- Intelligent traffic control strategies for VLC-connected vehicles and pedestrian flow managementPublication . Galvão, Gonçalo ; Vieira, Manuela ; Vieira, Manuel Augusto ; Véstias, Mário ; Louro, PaulaUrban traffic congestion leads to daily delays, driven by outdated, rigid control systems. As vehicle numbers grow, fixed-phase signals struggle to adapt to real-time conditions. This work presents a decentralized Multi-Agent Reinforcement Learning (MARL) system to manage a traffic cell composed of five intersections, introducing the novel Strategic Anti-Blocking Phase Adjustment (SAPA) module, developed to enable dynamic phase time adjustments. The goal is to optimize arterial traffic flow by adapting strategies to different traffic generation patterns, simulating priority movements along circular or radial arterials, such as inbound or outbound city flows. The system aims to manage diverse scenarios within a cell, with the long-term goal of scaling to city-wide networks. A Visible Light Communication (VLC) infrastructure is integrated to support real-time data exchange between vehicles and infrastructure, capturing vehicle position, speed, and pedestrian presence at intersections. The system is evaluated through multiple performance metrics, showing promising results: reduced vehicle queues and waiting times, increased average speeds, and improved pedestrian safety and overall flow management. These outcomes demonstrate the system’s potential to deliver adaptive, intelligent traffic control for complex urban environments.
- Enhanced random vector functional link networks with bayesian-based hyperparameter optimization for wind speed forecastingPublication . Seman, Laio Oriel ; Klaar, Anne Carolina Rodrigues ; Ribeiro, Matheus Henrique Dal Molin ; Stefenon, Stefano FrizzoAccurate short-term wind speed forecasting is essential for reliable and efficient wind energy integration. This paper introduces an enhanced Random Vector Functional Link (RVFL) network optimized through a Bayesian-based Neural Architecture Search (NAS) framework. The proposed RVFL-OptBayes model incorporates multi-scale feature generation, including kernel approximations, Nystr & ouml;m sampling, Fastfood transforms, wavelet scattering, and Neural Tangent Kernel embeddings with Principal Component Analysis (PCA)-aligned orthogonal initializations and spectral normalization to improve stability and feature diversity. Experiments were conducted on real-world Brazilian wind farm data to evaluate forecasting performance. Results show that RVFL-OptBayes outperforms conventional RVFL networks, deep learning models, and ensemble methods, achieving an R2 above 0.99. The proposed framework demonstrates that lightweight randomized architectures, when combined with principled hyperparameter search, can rival or surpass complex deep learning models for time-series forecasting. The findings suggest strong potential for practical deployment in renewable energy systems, offering accurate and computationally efficient wind speed predictions to support operational planning, grid stability, and smart energy management.
- Spatiotemporal wind energy forecasting: a comprehensive survey and a deep equilibrium-based case study with stemGNNPublication . Aquino, Luiza Scapinello; Seman, Laio Oriel; Mariani, Viviana Cocco; Coelho, Leandro Dos Santos; Stefenon, Stefano Frizzo; González, Gabriel VillarrubiaAccurate spatiotemporal wind energy forecasting is essential for ensuring grid stability and maximizing the efficiency of renewable energy systems. This paper addresses the challenge of modeling the complex spatial and temporal dependencies inherent in wind power generation by presenting a comprehensive survey of existing spatiotemporal forecasting methods and introducing an innovative deep learning approach. The proposed model integrates a Graph Neural Network (GNN) to represent wind turbines as nodes within a graph, capturing spatial relationships, while a Deep Equilibrium Model (DEQ) enables equilibrium-based inference to handle highly nonlinear wind patterns. A Sequence-to-Sequence (Seq2Seq) architecture further manages temporal dependencies. The method was validated using a real-world dataset of wind power generation, outperforming baseline models across multiple forecast horizons and maintaining stable accuracy across short- and mid-term predictions. Results demonstrate that the proposed GNN with DEQ effectively models both spatial and temporal dynamics for Seq2Seq data, improving prediction accuracy while maintaining computational efficiency. This study highlights the potential of equilibrium-based spatiotemporal graph models for wind energy forecasting and provides a robust tool for better integration of wind power into modern power grids.
- Multi-step short-term solar energy forecasting using Fourier-enhanced BiLSTM and neural additive modelsPublication . Seman, Laio Oriel; Stefenon, Stefano Frizzo; Yow, Kin-Choong; Coelho, Leandro dos Santos; Mariani, Viviana CoccoAccurate short and medium-term forecasting is important for mitigating uncertainty and enabling efficient energy grid management. While traditional machine learning and deep learning models offer improved accuracy, they often lack interpretability. To address these limitations, this study proposes a hybrid forecasting framework, called FNO-BiLSTM-NAM, that combines a Fourier Neural Operator (FNO) to extract spectral–temporal features, a Bidirectional Long Short-Term Memory (BiLSTM) network to model sequential dependencies, and a Neural Additive Model (NAM) to quantify feature-wise contributions. The model incorporates multi-scenario forecasting to support energy operators under different uncertainty levels. Experiments conducted on a dataset from a 5 MW PhotoVoltaic (PV) plant demonstrate the superiority of the model. For a 6-hour forecast horizon, the proposed FNO-BiLSTM-NAM model achieved a mean absolute error of 0.0712 and mean squared error of 0.0092, outperforming benchmark models across short- to medium-term horizons. Furthermore, the spectral analysis of the FNO revealed low-pass filtering behavior, highlighting the ability of the model to suppress high-frequency noise. Comparative experiments with five machine and deep learning baseline models confirm the robustness and generalization capacity of the framework. These results underscore the potential of the proposed model for enhancing PV energy forecasting accuracy while maintaining transparency across dynamic operating conditions.
- Novelty detection algorithms to help identify abnormal activities in the daily lives of elderly peoplePublication . Fernandes, Anita Maria da Rocha; Leithardt, Valderi Reis Quietinho; Santana, Juan F. de PazThe populations life expectancy is increasing, and this scenario will bring challenges to be faced in the coming decades to provide healthy and inclusive aging. At this stage of life, several common health conditions, chronic illnesses, and disabilities affect the individuals physical and mental health and prevent him from carrying out Activities of Daily Living. In this context, this article presents a comparative study between some Machine Learning algorithms used to identify behavioral abnormalities based on ADL (Activities of Daily Living), through the Novelty Detection technique. ADL data were used to create a model that defines the baseline behavior of an elderly person, and new observations, to verify significant changes in behavior, are classified as outliers or abnormal. The Local Outlier Factor, One-class Support Vector Machine, Robust Covariance, and Isolation Forest algorithms were analyzed, and the Local Outlier Factor obtained the best result, reaching a precision and F1-Score of 96%.