Repositório UFSCar

O Repositório Institucional da UFSCar (RI UFSCar) é um sistema de informação que visa armazenar, preservar, organizar e disseminar amplamente a produção intelectual dos diversos setores e segmentos da comunidade da UFSCar, provendo o acesso aberto à informação produzida na instituição e registrada como científica, tecnológica, didática, artístico-cultural e técnico-administrativa.

Para mais informações, acesse: RI UFSCar

Instruções aos autores: Clique aqui.

Comunidades

Selecione uma comunidade para navegar por suas coleções

Agora exibindo 1 - 5 de 6

Submissões Recentes

Simulador educacional de NoC: visualização do funcionamento da rede e estudo de algoritmos de mapeamento
(Universidade Federal de São Carlos, 2026-06-26) Santos, Breno Dias Arantes dos; Pedrino, Emerson Carlos; http://lattes.cnpq.br/6481363465527189
In the literature, several simulators for Many-Core processors and Network-on-Chip (NoC) architectures have been proposed. However, most of these tools present a high level of complexity, requiring advanced knowledge of low-level hardware concepts, which makes them unsuitable for undergraduate students, especially in early stages. Furthermore, there is a lack of accessible simulators developed in high-level programming languages such as Python. This work presents the development of an educational simulator for Many-Core systems based on NoC architectures, focusing on accessibility and usability without compromising the accuracy of performance analysis. The proposed simulator abstracts key structural and functional aspects of NoCs, including topology configuration, packet management, routing algorithms, and performance metrics such as latency, energy consumption, and fault tolerance. Additionally, the tool provides support for task mapping algorithms and enables intuitive interaction through a graphical interface, allowing users to visualize system behavior dynamically. The simulator was validated by comparing its results with baseline studies from the literature, demonstrating similar performance trends. The proposed approach contributes to the democratization of access to NoC simulation tools, offering an effective platform for teaching and experimentation in undergraduate Computer Engineering courses.
Estudo de efeito da turbulência em uma placa plana com cavidade usando o programa Ansys student
(Universidade Federal de São Carlos, 2026-07-13) Ferrari, Daniel Ferezin; Medeiros, Marcello Augusto Faraco de; http://lattes.cnpq.br/7772253950731916; https://orcid.org/0000-0003-4174-6258; Campo Júnior, Vivaldo Leiria; http://lattes.cnpq.br/7150892281897003; https://orcid.org/0000-0002-1250-1983
Turbulence is a undesired effect that happens in aeronautics. With greater turbulence comes also a greater drag, which makes the airplane to burn more fuel. Not only that, but computational simulations do not take into account minor imperfections that occur during the aircraft manufacturing, generating discrepancies between theoretical simulations and practical results, resulting in manufacturing having big error margins. Whit that in mind, the Group of Acoustics, Turbulence and Transition (GATT) from EESC - USP Sao Carlos initiated studies about the effects of a gap in a wing. Due to the particularity of such case, GATT have done computational studies based on models developed by GATT itself. To provide an independent theoretical result, this work tries to accomplish a simulation of the same case using a commercial software available for free, ANSYS Student, to verify if the computational results agree with the experimental ones obtained by GATT.
Reconhecimento de expressões faciais usando deep learning: uma análise comparativa sistemática em cenários de mundo real
(Universidade Federal de São Carlos, 2026-09-04) Pacini, João Victor; Levada, Alexandre Luis Magalhães; http://lattes.cnpq.br/3341441596395463
This work presents a systematic investigation into the performance and robustness of deep learning architectures applied to Facial Expression Recognition (FER) in real-world scenarios. The central motivation lies in the importance of emotion recognition as a fundamental component of human-computer interaction, with applications ranging from telehealth to social robotics, focusing on performance in real-world scenarios. The main objective is to analyze the impact of the domain shift phenomenon on the transition from controlled laboratory data to dynamic real-world environments, as well as to mitigate the effects of occlusions and class imbalance. To this end, a comparative evaluation of spatial feature extraction models was conducted, including the ResNet50V2, EfficientNetB0, and ConvNeXt Tiny networks, as well as a temporal hybrid model composed of a ConvNeXt architecture integrated with an LSTM recurrent network equipped with attention mechanisms. Spatial modeling experiments revealed the superiority of the ConvNeXt Tiny model pre-trained with large datasets, which achieved an overall accuracy of 87.16% on the RAF-DB dataset, highlighting the crucial role of the statistical normalization step and sequential pre-training. In contrast, cross-validation tests revealed a drastic 53% drop in model performance when exposed to domain transfer between RAF-DB and CK+ (with accuracy declining to 34.05%), highlighting the conceptual fragility in the face of variations in poses and occlusions. In temporal experiments with the CK+ dataset, the hybrid architecture demonstrated high performance under strict subject-independent validation, reaching 95.14% overall accuracy when tracking facial muscle dynamics to the peak of expression. Additionally, ablation studies focused on class imbalance indicated that simple weighting methods in the loss function (Class Weights) outperformed synthetic interpolation strategies in latent space (SMOTE) and pixel space (MixUp), preventing overfitting and preserving subtle facial distortions. It is concluded that, although modern deep architectures exhibit high hierarchical representation capabilities, practical generalization is still severely limited by domain shift, highlighting the future relevance of domain adaptation and explainability techniques.
Identidades polinomiais para a álgebra de Grassmann
(Universidade Federal de São Carlos, 2026-07-16) Grytz, Vítor do Vale; Fagundes, Pedro Souza; http://lattes.cnpq.br/3700587815257392
In this work, we study algebras with polynomial identities (PI-algebras), with particular emphasis on the polynomial identities of the Grassmann algebra. To this end, we develop the necessary background on algebras, free algebras, and T-ideals. In addition to the polynomial identities of the Grassmann algebra, we also present the polynomial identities of the matrix algebras M2(K) and Un(K). Furthermore, we study the T-ideal of commutative algebras and, finally, present a basis for the T-ideal of the Grassmann algebra.
Aesthetics punk em discurso: efeitos de sentido no Instagram sob a perspectiva da Análise do Discurso Digital
(Universidade Federal de São Carlos, 2026-07-17) Tonhatti Ciniello, Giulia; Fernandes Araújo , Dandara Rochelly; http://lattes.cnpq.br/3509447023315693; https://orcid.org/0009-0006-1235-2988; Nogueira, Luciana; https://lattes.cnpq.br/7475342844385424; https://orcid.org/0000-0003-4470-5005; http://lattes.cnpq.br/4181355889127433; https://orcid.org/0009-0004-9089-1800
Currently, "aesthetics" are considered a phenomenon in digital culture that fosters a sense of identification between individuals and visual elements. Thus, this research aims to analyze the meaning effects produced by the circulation of the "punk aesthetic" on Instagram, drawing on Digital Discourse Analysis — specifically the formulations of Marie-Anne Paveau (2021) and Cristiane Dias (2016) — while considering its modes of constitution, formulation, and circulation (Orlandi, 2001), as well as the modes of subject constitution within these spaces. To this end, concepts from Materialist Discourse Analysis and Digital Discourse Analysis are employed, including conditions of production, discursive memory, metallic memory, techno-discursive memory, and the archive, as defined by Orlandi (2015), Leandro-Ferreira (2020), and Paveau (2021). Specifically, we analyze excerpts from Instagram posts related to the punk aesthetic, seeking to understand the meaning effects produced when this aesthetic circulates as a techno-discourse. The corpus was compiled from Instagram posts and materials from the Aesthetics Wiki, a collaborative platform dedicated to cataloging and describing aesthetics. Methodologically, the research is grounded in the notion of the "excerpt" proposed by Orlandi (1984), viewing the corpus as a set of interconnected discursive materialities interpreted and analyzed through the theoretical framework of Discourse Analysis. Through this work, we hope to contribute to discussions regarding the production of meaning associated with aesthetics in digital spaces and to reflections on subject constitution within contemporary digital materialities.