http://repositorio.ufc.br/handle/riufc/478 2026-06-21T23:33:22Z http://repositorio.ufc.br/handle/riufc/86856 Título: Dispositivos lógicos lineares totalmente ópticos em acopladores de fibra de cristal fotônico de três núcleos Autor(es): Rodrigues, João Paulo Teófilo Abstract: The growth in global data traffic and the demand for energy efficiency in Artificial Intelligence training have driven the search for alternatives that overcome the limitations of conventional electronic processing. This reality has led the computer engineering community to revisit paradigms established in recent decades. Traditionally, the premise has been consolidated that all-optical processing strictly requires highly nonlinear media (such as the Kerr effect) or the mediation of optoelectronic devices with successive electro-optic conversions. This work demonstrates a paradigm shift by presenting the analytical design and numerical validation of two all-optical logic devices operating strictly in the linear regime: a multifunctional logic gate and a configurable multifunctional logic device. In both architectures, the fundamental Boolean operations OR and AND are dynamically enabled based on the logic levels input into a selector channel. Each component is structured from a single Photonic Crystal Fiber (PCF) segment composed of three coupled cores, arranged in planar and triangular layouts, respectively. Robust mathematical modeling demonstrates that the control and routing of light states occur purely in the linear regime through coupling guided by the fiber’s geometric design, enabling the processing of ultrashort, low-power pulses with a high contrast ratio via Pulse Amplitude Modulation (PAM). Interconnecting the fundamentals of Applied Electromagnetics and Digital Systems Engineering, this thesis establishes the microstructured design of functional fibers as a viable, high-performance platform for native optical computing. Tipo: Tese 2025-01-01T00:00:00Z http://repositorio.ufc.br/handle/riufc/86656 Título: Métodos para treinamento rápido e esparso de máquinas de vetores-suporte de mínimos quadrados: uma abordagem dual Autor(es): Marinho, Felipe Pinto Abstract: The least squares support vector machine model is a variant of the classical support vector machine model that employs equality constraints in the formulation of its primal problem. This allows the derivation of a linear system when applying the Karush–Kuhn–Tucker optimality conditions, considerably simplifying the training of this model when compared to the adjustment of support vector machines. However, a drawback of this formulation lies in the fact that the optimal vector of Lagrange multipliers of the problem is dense. Thus, all training patterns are considered support vectors, making the prediction stage computationally expensive when working with large datasets. In many cases, the solution of the system is obtained through the use of iterative methods based on conjugate directions, which, on the one hand, is advantageous since it avoids numerical difficulties related to matrix inversion, but, on the other hand, makes the training stage slow for datasets with high volume, as it is necessary to operate with dense kernel matrices. In this context, two new methodologies are proposed for the fast and sparse training of least squares support vector machines. In the first approach, the dual problem of least squares support vector machines is solved via a sequential minimal optimization algorithm with a new three-term conjugate descent direction which, combined with a working set selection strategy based on functional gain, allows an acceleration in convergence, reducing the number of iterations when compared to the standard sequential minimal optimization algorithm. In addition, an iterative pruning process based on the functional gain of the optimization problem is adopted in order to sparsify the obtained Lagrange multipliers. Finally, the last proposal consists of the use of a new spectral conjugate gradient method for solving the corresponding dual problem and sparsification through iterative pruning using the proximity of the pattern to the decision hyperplane as the criterion for removal. Numerical experiments carried out on several real and artificial datasets demonstrate that both approaches present competitive performance, with fast training and a high level of sparsity of the Lagrange multipliers. For binary classification datasets, the sparsity gain reached approximately 80% when compared to the total number of training samples for the considered dataset. The reduction in training time was approximately 99.9% in relation to standard least squares support vector machines. For datasets with higher volume, the proposals were the only ones that provided feasible training time with stable convergence. The quality of the decision boundaries was further analyzed for synthetic datasets, where the results indicate the generation of boundaries similar to the considered benchmarking model, confirming the predictive capability of the new methodologies. Finally, the results for regression datasets indicate that the proposal based on the spectral conjugate gradient method may be a sparse and fast-training alternative to the least squares support vector machine regression model. Tipo: Tese 2025-01-01T00:00:00Z http://repositorio.ufc.br/handle/riufc/86609 Título: Análise de dados de vibração não-estacionários em aerogeradores usando algoritmos de aprendizagem profunda potencializados por estimadores de densidade kernel Autor(es): Nogueira, Tiago de Oliveira Abstract: Wind turbines are essential components for incorporating sustainable measures in energy production. However, the effectiveness of condition monitoring systems (CMS) is often challenged by stochastic and broadband noise, which can mask harmonic fault components in vibration signals. Seeking to solve this problem, signal filtering methods can be used in conjunction with multiple approaches to pattern recognition techniques. Classical machine learning models (KNN, SVM, ANN) can be used to evaluate the filtering impact on sets of statistical features extracted from the spectrum. In parallel, deep learning techniques, specifically Convolutional Neural Networks (CNNs), can be applied for end-to-end classification directly from spectrograms. This thesis proposes and validates a new spectral preprocessing methodology, based on Kernel Density Estimation (KDE), to selectively remove noise and optimize the separability of fault features. The validation was structured in two phases: a Phase 1 of controlled validation on the VBL database (comparing KDE, DWT, and raw data) and a Phase 2 of real-world application on the EISLAB database (anomaly detection in wind turbines). Both training paradigms were explored: Supervised Learning, where the model learns to map inputs to known fault labels, and Self-Supervised Learning (via SimCLR), focused on learning robust data representations without using labels. The results from Phase 1 (VBL) demonstrated the effectiveness of KDE in optimizing robust models (SVM/ANN) and, more expressively, in CNN models, where the KDE + Supervised CNN combination achieved the highest validation accuracy (98.5%), surpassing the raw data (87.0%) and the DWT (96.0%). In Phase 2 (EISLAB), the application of the KDE filter to real-world wind turbine data again proved superior, increasing the Supervised CNN accuracy from 68.5% to 72.3% and, crucially, raising the F1-Score for the fault class from 0.554 to 0.600. It is concluded that adaptive filtering via KDE is an effective methodological tool, acting as a spectral feature optimizer, significantly increasing the precision and robustness of artificial intelligence-based diagnostic systems. The experiments showed that KDE significantly improves robust models, especially CNNs, achieving the best results both in a controlled environment and with real-world wind turbine data. Thus, the method stands out not only as a filter but as a spectral feature optimizer, enhancing accuracy and reliability in fault diagnosis for rotating systems. Tipo: Tese 2026-01-01T00:00:00Z http://repositorio.ufc.br/handle/riufc/86147 Título: Detection and characterization of equatorial ionospheric scintillation based on GNSS observations using convolutional neural networks Autor(es): Pacelli, Rubem Vasconcelos Abstract: Global Navigation Satellite Systems (GNSS) receivers operating in the equatorial region are severely affected by ionospheric scintillation caused by rapid fluctuations in the amplitude and phase of the received electromagnetic wave. To monitor scintillation activity, commercial receivers typically provide the S4 and 𝛔ɸ indices, which are computed using ad-hoc thresholds and detrending techniques. In addition to the intrinsic limitation of being restricted to an M-ary classification problem, these filtering methods are known to mask scintillation events. The present thesis proposes a paradigm shift in the characterization of ionospheric scintillation. Instead of relying on detrending techniques and preset thresholds, recent Bayesian methods are employed to disentangle the scintillation signal from line-of-sight dynamics, and new convolutional neural network algorithms are investigated for the characterization task. A simplified signal model is derived and used to generate synthetic scintillation time series based on two ionospheric scintillation models: the Cornell Scintillation Model and the compact phase-screen-based scintillation model. The main contribution of this thesis is a new framework that combines convolutional neural networks and class activation maps as a post-hoc interpretability tool that highlights input regions most relevant to the discriminative task. This approach makes it possible to analyze samples whose features are prone to be classified as strong rather than weak scintillation, and the converse, without setting arbitrary thresholds or relying on ad-hoc heuristics established to obtain scintillation indices. The results demonstrate an average accuracy above 89%, which enables sample-level characterization of the severity of scintillation and provides significant improvements compared to traditional threshold-based classification methods used in ionospheric scintillation monitoring applications. Tipo: Tese 2026-01-01T00:00:00Z