Doutorado em Engenharia Elétrica

URI Permanente para esta coleção

http://repositorio.ufes.br/handle/10/17618

Nível: Doutorado
Ano de início:
Conceito atual na CAPES:
Ato normativo:
Periodicidade de seleção:
Área(s) de concentração:
Url do curso:

Navegar

Agora exibindo 1 - 5 de 140

A socially assistive robot as a therapeutic tool for applied behavior analysis therapy in children with autism spectrum disorder through dynamically modulated serious games
(Universidade Federal do Espírito Santo, 2024-10-07) Freitas, Éberte Valter da Silva; Bastos Filho, Teodiano Freire; https://orcid.org/0000-0002-1185-2773; Caldeira, Eliete Maria; Cuadros, Marco Antonio de Souza Leite; Valadão, Carlos Torturella; Vasquez, Luiz Fernando Guerrero
This work presents the programming/adaptation of the Socially Assistive Robot (SAR) called Mobile Autonomous Robot for Interaction with Autistics and Trisomy 21 (MARIA T21) to be able to carry out interventions, according to the Applied Behavior Analysis (ABA) therapy, applied to Children with Autism Spectrum Disorder (ASD), Serious Games (SGs), which are projected by MARIA T21, with difficulty modulation dynamics, are used as part of this therapy thus creating a motivating and facilitating effect for children and therapists. The SGs were developed in Unity 3D, using C Sharp language (C#), and modulated according to the child’s performance, eye attention level, and user’s facial emotion, being integrated into the robot through the Robot Operating System (ROS). The child-Robot Interaction (CRI) protocol followed in this research was carried out in a child’s psychotherapy room at the APAE (Association of Parents and Friends of Excepcional People for the acronym in Portuguese) of Vitoria in Espirito Santo state (Brazil), which was instrumented with four video cameras and supervised by a group of researchers. The sample consisted of 18 children with a conclusive diagnosis of ASD, 3 girls and 15 boys, aged 5 to 9 years and presence of stereotyped movements of eyes and/or hands and/or feet. The experiments were separated into three modules for which SGs or specific therapeutic dynamics were applied, namely: Cognitive Module, Physical Module and Functional Module. The first one brings together the games and dynamics applied by MARIA T21 focused on the development and improvement of cognitive learning skills, whereas the physical module brings SGs for motor evaluation and correction, that can be used for physical and postural strengthening by a physiotherapist. Finally, the functional module has applications of SGs and Occupational Therapy dynamics in order to work on the so-called Activities of Daily Living (ADLs) and encourage autonomy in those assisted. In addition to the data recorded by MARIA T21, two evaluation scales– Pediatric Evaluation of Disability Inventory Computer Adaptive Test (PEDI-CAT) and System Usability Scale (SUS)– were applied for each module, and another for the degree of acceptance of the robot in therapy by the child. The results obtained so far enable the use MARIA T21 as an ABA therapeutic tool. In addition, the SGs are capable of dynamically modulating their difficulty, providing greater user adherence and continued attention in the optimal learning zone of the ABA bibliography while carrying out the activities. The technology embedded in the robot has also enabled the identification and quantification of characteristics and parameters, such as the presence and recurrence of stereotypies and postural dysregulations, placing the robot as an innovative and promising tool to assist health professionals in the early diagnosis, conduction and follow-up of therapies.
Extracting pulse rate, oxygen saturation, and respiration rate through smartphones
(Universidade Federal do Espírito Santo, 2024-08-14) Lampier, Lucas Côgo; Co-orientador1; https://orcid.org/; http://lattes.cnpq.br/; Co-orientador2; https://orcid.org/; http://lattes.cnpq.br/; Co-orientador3; https://orcid.org/; http://lattes.cnpq.br/; Co-orientador4; ID do co-orientador4; Lattes do co-orientador4; Bastos Filho, Teodiano Freire ; https://orcid.org/; http://lattes.cnpq.br/; Orientador2; https://orcid.org/; http://lattes.cnpq.br/; https://orcid.org/; http://lattes.cnpq.br/; Krishnan, Sridhar; https://orcid.org/; http://lattes.cnpq.br/; Andrade, Adriano de Oliveira; https://orcid.org/; http://lattes.cnpq.br/; Coelho, Yves Luduvico; https://orcid.org/; http://lattes.cnpq.br/; Floriano, Alan Silva da Paz; http://lattes.cnpq.br/; Ciarelli, Patrick Marques; https://orcid.org/; http://lattes.cnpq.br/; Caldeira, Eliete Maria de Oliveira; https://orcid.org/; http://lattes.cnpq.br/; 7º membro da banca; https://orcid.org/; http://lattes.cnpq.br/
In the last years, the power of smartphones has been increasing. These devices, equipped with multiple sensors and a high computational power, have become an essential part of daily life. With their increasing capabilities, smartphones are no longer limited to basic functions, but have emerged as versatile tools that can be utilized for multiple healthcare purposes. This work aims to use cameras to extract pulse rate and oxygen saturation, and microphones to measure respiration rate. Multiple methods to measure pulse rate, oxygen saturation and respiration rate using a color camera and a microphone are evaluated to be applied to the smartphone. New methodologies based on Deep Learning (DL) to infer pulse rate and oxygen saturation of people using a color camera are also presented, and a methodology to extract respiration rate using a smartphone microphone is also evaluated. It is shown that the DL models proposed are more accurate in measuring oxygen saturation and pulse rate from small length signals than conventional methods proposed in the literature. Using these model, the Root Mean Squared Error (RMSE) of the oxygen saturation model was 2.92%, and the Spearman Rank Correlation Coefficient (SRCC) was 0.95. The pulse rate was measured remotely and with the skin in contact with the camera. When the skin is contact with the camera, the pulse rate RMSE was 1.78 BPM and an SRCC of 0.96. When the pulse rate was measured remotely, the RMSE was 3.93 BPM and the SRCC was 0.86. The respiration rate method also presented a low error, with RMSE of 0.74 breaths/min and a SRCC of 0.99. Finally, a prototype of an Android application compiling the techniques to measure oxygen saturation, pulse rate, and respiration rate was built. The application was tested with eight volunteers, and the results showed that the pulse rate and respiration rate presented low error, RMSE of 4.54 BPM and 0.74 breaths/min, respectively. However, the oxygen saturation model did not perform well in the application (RMSE of 4.37%), most likely due to the differences between the setups used to record the model’s training images, and to collect data using the application
FBG-based sensors for oil and gas industry: assessment of heat transfer, structural health, liquid level, thermal conductivity and salinity
(Universidade Federal do Espírito Santo, 2024-06-03) Lazaro, Renan Costa; Frizera Neto, Anselmo; https://orcid.org/0000-0002-0687-3967; http://lattes.cnpq.br/; Leal Junior, Arnaldo Gomes; https://orcid.org/0000-0002-9075-0619; http://lattes.cnpq.br/; Orientador2; https://orcid.org/; http://lattes.cnpq.br/; https://orcid.org/; http://lattes.cnpq.br/; Pontes, Maria José; https://orcid.org/0000-0002-9009-2425; http://lattes.cnpq.br/; Marques, Carlos Alberto Ferreira; https://orcid.org/; http://lattes.cnpq.br/; Theodosiou, Antreas; https://orcid.org/; http://lattes.cnpq.br/
This doctoral thesis focuses on the advancement of optical fiber-based sensors employing Fiber Bragg Gratings (FBG) for enhanced sensing in the oil and gas industry. The primary aim is to refine the evaluation of thermophysical parameters of fluids in classified and flammable environments. The research introduces FBG-based for tank structural health monitoring and parameters measurements in fluids, such as temperature, level, thermal conductivity and salinity. Experimental results demonstrate the efficacy of these sensors in challenging industrial conditions. The thermal experiments, utilizing an FBG-based temperature sensor, reveal insights into thermal power distribution in liquid processing systems. Specific heat and thermal conductivity of water are successfully estimated, demonstrating increased thermal stability with higher heat power. A method for measuring heat transfer rate in liquids is proposed, showing potential applications in industrial contexts. In the realm of structural health monitoring (SHM), the quasi-distributed FBG sensor, combined with supervised machine learning, exhibits high accuracy in monitoring stress and deformation in oil tank structures. The Random Forest algorithm enables precise liquid level estimation with minimal error, contributing to predictive maintenance strategies. The development of an all-optical hot-wire sensor showcases its precision in assessing thermal conductivity and salinity in various fluids. The sensor, integrating FBG with a hot-wire component, proves effective in discriminating substances with close thermal conductivity values. Future work aims at reducing measurement times and adapting the sensor for direct salinity measurement. Finally, worth highlighting the significant contributions of each sensor, emphasizing the practical applicability and promising results obtained in thermal analysis, structural health monitoring and all-optical sensing for the oil and gas industry. The research sets the stage for further exploration and refinement of these sensor technologies in complex industrial scenarios.
Methodologies to improve one-class classifier performance applied to multivariate time series
(Universidade Federal do Espírito Santo, 2024-04-05) Machado, André Paulo Ferreira; Ciarelli, Patrick Marques ; https://orcid.org/; http://lattes.cnpq.br/; Co-orientador2; https://orcid.org/; http://lattes.cnpq.br/; Co-orientador3; https://orcid.org/; http://lattes.cnpq.br/; Co-orientador4; ID do co-orientador4; Lattes do co-orientador4; Munaro, Celso José ; https://orcid.org/; http://lattes.cnpq.br/; Orientador2; https://orcid.org/; http://lattes.cnpq.br/; https://orcid.org/; http://lattes.cnpq.br/; Vargas, Ricardo Emanuel Vaz ; https://orcid.org/; http://lattes.cnpq.br/; Serra, Ginalber Luiz de Oliveira ; https://orcid.org/; http://lattes.cnpq.br/; Coelho, Leandro dos Santos ; https://orcid.org/; http://lattes.cnpq.br/; Lima Netto, Sergio ; http://lattes.cnpq.br/; 5º membro da banca; https://orcid.org/; http://lattes.cnpq.br/; 6º membro da banca; https://orcid.org/; http://lattes.cnpq.br/; 7º membro da banca; https://orcid.org/; http://lattes.cnpq.br/
This work proposes novel methodologies to improve the performance of one-class classifiers applied to multivariate time series data. The main method is through clustering of multivariate time series. Datasets arising from real processes come from the available sensors and are affected by many factors, such as aging of the process, changes in the operation region, and equipment malfunction. Despite that, one expects that the classes represented by such diverse data can be unveiled via trained classifiers. This work hypothesizes that the overall performance can be improved by training sets of one-class classifiers with subsets of data clustered by similarity, obtained by DTW Barycenter Averaging (DBA) which is used to measure the similarity between the time series and each cluster. The proposed method is applied to one class classifiers since they are trained only with the target class, which is clustered based on time series similarity using Dynamic Time Warping and k-means. Additionally, a second approach is proposed, called time-shift of labels, to improve the differentiation between normal and faulty data. This method is applied during the training phase and focuses on particular situations involving the transition from normality to faulty data, where the boundaries are difficult to differentiate (overlapping data). The time-shift results show a mitigation of the effect of overlapping data. The advantages of the techniques are illustrated through their application to two public datasets one from the oil industry with instances characterizing eight classes of data represented by five time series (3W dataset), and another from a hydraulic system for the study of typical hydraulic system failures with five classes and seventeen time series (Condition monitoring of hydraulic systems - ICM dataset). For the 3W dataset, seven classes are selected to train Long Short Term Memory (LSTM) classifiers using the variables and instances clustered using time series clustering algorithms. The results demonstrate that increasing the similarity of training data tends to improve the performance of the LSTM classifier, achieving an increase of 10% in the overall performance on the 3W dataset. In a specific case, where the clustering model raised the similarity by 84%, the classification performance improved by 21%. For condition monitoring of hydraulic system data, the proposed method achieved a significant performance improvement of over 40% compared to the baseline model. Notably, in the specific case of leakage fault, the classification performance improvement rises by 64%
Detecção de eventos de obstrução em válvula no processo de lingotamento contínuo com uso de aprendizado de máquinas
(Universidade Federal do Espírito Santo, 2024-05-25) Diniz, Ana Paula Miranda; Co-orientador1; https://orcid.org/; http://lattes.cnpq.br/; Co-orientador2; https://orcid.org/; http://lattes.cnpq.br/; Co-orientador3; https://orcid.org/; http://lattes.cnpq.br/; Co-orientador4; ID do co-orientador4; Lattes do co-orientador4; Orientador1; https://orcid.org/; http://lattes.cnpq.br/; Orientador2; https://orcid.org/; http://lattes.cnpq.br/; https://orcid.org/; http://lattes.cnpq.br/; 1º membro da banca; https://orcid.org/; http://lattes.cnpq.br/; 2º membro da banca; https://orcid.org/; http://lattes.cnpq.br/; 3º membro da banca; https://orcid.org/; http://lattes.cnpq.br/; 4º membro da banca; http://lattes.cnpq.br/; 5º membro da banca; https://orcid.org/; http://lattes.cnpq.br/; 6º membro da banca; https://orcid.org/; http://lattes.cnpq.br/; 7º membro da banca; https://orcid.org/; http://lattes.cnpq.br/
The continuous casting process, used in the manufacture of steel plates, is currently the most economical and efficient way of production within the industry. Although continuous casting is a widely used process, some problems associated with the process have not yet been resolved, one of them being the obstruction of the Submerged Entry Nozzle (SEN), which controls the flow of steel between the tundish and the mold. This obstruction, also called clogging, not only impairs the quality of the product but also results in lower process yield, resulting in losses. Thus, clogging detection is of fundamental importance, because control actions can allow the system to operate for a longer time. In this work, methodologies based on Machine Learning and Deep Learning are presented and compared to detect the occurrences of clogging from historical data of process variables. In general, the performance of the classifiers achieved very promising results in real data with unbalanced classes. In particular, the method employing spatiotemporal analysis, using four process variables, obtained a remarkably superior performance when compared to the others, reaching averages of Precision and Recall, respectively, of 95.53% and 97.33%. To reduce the false positive and negative rates, a post-processing heuristic was implemented and applied to the model output, achieving a Precision and a Recall, respectively, of approximately 98% and 99%. To the best of our knowledge, these results have never been found in the literature. Although a detailed comparison is unfeasible due to the differences between the datasets and their inaccessibility, the modeling proposed here reached higher performance levels when compared to the results found in the literature to solve this industry’s problem. The high and unprecedented results obtained in this work, therefore, will contribute both to the improvement of the quality of the final product and to the reduction of costs associated with steel production, since the more effective classification of clogging occurrences can help operators in the corrective action planning

Navegar

Submissões Recentes