Programa de Pós-graduação em Engenharia Elétrica e da Computação

URI Permanente desta comunidade

http://200.137.215.59/tede/handle/tde/262

Navegar

Navegando Programa de Pós-graduação em Engenharia Elétrica e da Computação por Por Orientador "Cruz Junior, Gelson da"

Agora exibindo 1 - 6 de 6
  • Nenhuma Miniatura disponível
    Item
    Neuroevolucão de um controlador neural e dinâmico para um robô móvel omnidirecional de quatro rodas
    (Universidade Federal de Goiás, 2018-11-01) Domingos, Ruan Michel Martins; Cruz Junior, Gelson da; http://lattes.cnpq.br/4370555454162131; Cruz Junior, Gelson da; Soares, Fabrizzio Alphonsus Alves de Melo Nunes; Brito, Leonardo da Cunha; Vinhal, Cássio Dener Noronha
    This work proposes a hierarchical control architecture to deal with the Trajectory Tracking Problem while an autonomous omnidirectional wheeled mobile robot operates. A traditional velocity controller and an intelligent decision-making neural network controller address the problem, considering the robot's kinematic and dynamic models. A neuroevolution technique evolves a smart Neurocontroller functionally attached to a Resolved Acceleration PI/PD Controller. The resulting control strategy shows to improve trajectory tracking errors during simulation studies. The Traditional and Intelligent controller combination showed very promising results even when applied in other trajectories that didn't belong to the original training set.
  • Nenhuma Miniatura disponível
    Item
    Comparação de algoritmos de enxame de partículas para otimização de problemas em larga escala
    (Universidade Federal de Goiás, 2018-10-26) Melo, Leonardo Alves Moreira de; Cruz Junior, Gelson da; http://lattes.cnpq.br/4370555454162131; Silva, Karina Rocha Gomes da; Rodrigues, Cássio Leonardo; Cruz Junior, Gelson da
    In order to address an issue concerning the increasing number of algorithms based on particle swarm optimization (PSO) applied to solve large-scale optimization problems (up to 2000 variables), this article presents analysis and comparisons among five state- of-the-art PSO algorithms (CCPSO2, LSS- PSO, OBL-PSO, SPSO and VCPSO). Tests were performed to illustrate the e ciency and feasibility of using the algorithms for this type of problem. Six benchmark functions most commonly used in the literature (Ackley 1, Griewank, Rastrigin, Rosenbrock, Schwefel 1.2 and Sphere) were tested. The experiments were performed using a high-dimensional problem (500 variables), varying the number of particles (50, 100 and 200 particles) in each algorithm, thus increasing the computational complexity. The analysis showed that the CCPSO2 and OBL-PSO algorithms found significantly better solutions than the other algorithms for more complex multimodal problems (which most resemble realworld problems). However, considering unimodal functions, the CCPSO2 algorithm stood out before the others. Our results and experimental analysis suggest that CCPSO2 and OBL- PSO seem to be highly competitive optimization algorithms to solve complex and multimodal optimization problems.
  • Nenhuma Miniatura disponível
    Item
    Modelo neural recozido para a representação semântica de documentos por meio de vetores contínuos
    (Universidade Federal de Goiás, 2020-11-13) Mendonça, Leandro Rezende Carneiro de; Cruz Junior, Gelson da; http://lattes.cnpq.br/4370555454162131; Cruz Junior , Gelson da; Soares Alcalá , Symone Gomes; Oliveira , Marco Antonio Assfalk de; Soares , Fabrízzio Alphonsus Alves de Melo Nunes; Campos , Sérgio Vale Aguiar
    As a result of the growing production of unstructured textual data, techniques for representing words and documents in the vector space have emerged recently. The Brazilian Public Ministry has received several textual requests that are send by citizens with different needs, such as those involved in cases of domestic violence against women, others requesting intensive care unit admissions, and more. The time spent in classifying, detecting similar requests and distributing them is essential to optimize and save public resources. Therefore, we adopted the neural model with the Simulated Annealing (SA), a classic global optimization algorithm with low computational complexity, because of the need to reduce the daily training time, providing a more friendly graphic visualization of data in high dimensions, supporting the judicial decision process. The physical analogy of the SA meta-heuristic associated with the continuous representation of documents in the vector space contribute greatly to the friendly visualization of a high-dimensional dataset, maintaining a comparable score with other deep models and optimization algorithms, such as Covariance Matrix Adaptation Evolution Strategy (CMA-ES) and Bayesian Optimization (BO).
  • Nenhuma Miniatura disponível
    Item
    Modelos de simulação e otimização para sistemas hidrotérmicos
    (Universidade Federal de Goiás, 2016-09-15) Ramos, Edson da Silva; Cruz Junior, Gelson da; http://lattes.cnpq.br/4370555454162131; Cruz Júnior, Gelson da; http://lattes.cnpq.br/4370555454162131; Soares, Fabrizzio Alphonsus Alves de Melo Nunes; Silva, Karina Rocha Gomes da
    The problem of planning the hydrothermal systems is complex, dynamic, stochastic, interconnected and nonlinear. In this work this problem is treated to meet one goal: minimize the use of water tank during a scenario of natural river flows lean period. This paper presents the application of meta-heuristics mono-objective of this problem, using a set of eight real plants in the National Interconnected System during the period of five years. The algorithms used were: PSO, ABeePSO, LSSPSO and KFPSO. The experiments were compared to studies using Nonlinear Programming and it appears that this work presents a simulation model and optimization for flexible hydrothermal system and highly adaptable to the use of different meta-heuristics allowing the researcher to apply different algorithms and compare the results between them.
  • Nenhuma Miniatura disponível
    Item
    Novas abordagens para segmentação de nuvens de pontos aplicadas à robótica autônoma e reconstrução 3D
    (Universidade Federal de Goiás, 2016-08-12) Santos, Gilberto Antônio Marcon dos; Vinhal, Cassio Dener Noronha; http://lattes.cnpq.br/9791117638583664; Cruz Junior, Gelson da; http://lattes.cnpq.br/4370555454162131; Cruz Junior, Gelson da; Bastos Filho, Carmelo Jose Albanez; Brito, Leonardo da Cunha; Vinhal, Cassio Dener Noronha
    Depth sensing methods yield point clouds that represent neighboring surfaces. Interpreting and extracting information from point clouds is an established field, full of yet unsolved challenges. Classic image processing algorithms are not applicable or must be adapted because the organized structure of 2D images is not available. This work presents three contribution to the field of point cloud processing and segmentation. These contributions are the results of investigations carried out at the Laboratory for Education and Innovation in Automation – LEIA, aiming to advance the knowledges related to applying spacial sensing to autonomous robotics. The first contribution consists of a new algorithm, based on evolutionary methods, for extracting planes from point clouds. Based on the method proposed by Bazargani, Mateus e Loja (2015), this contribution consists of adopting evolutionary strategies in place of genetic algorithms making the process less sensitive to user-defined parameters. The second contribution is a method for segmenting ground and obstacles from point clouds for autonomous navigation, that utilizes the proposed plane extraction algorithm. The use of a quadtree for adaptive area segmentation allows for classifying points with high accuracy efficiently and with a time performance compatible with low cost embedded devices. The third contribution is a variant of the proposed segmentation method that is more noise tolerant and robust by incorporating a neural classifier. The use of a neural classifier in place of simple thresholding makes the process less sensitive to point cloud noise and faults, making it specially interesting for processing point clouds obtained from real time stereo reconstruction methods. A through sensitivity, accuracy, and efficiency analysis is presented for each algorithm. The dihedral angle metric (angle between the detected plane and the reference polygons that share at least one point) proposed by Bazargani, Mateus e Loja (2015) is used to quantify the plane detection method accuracy. The ratio between the correctly classified points and the total number of points is utilized as an accuracy metric for the ground segmentation methods. Additionally, computing costs and execution times are considered and compared to the main state-of-the-art methods.
  • Nenhuma Miniatura disponível
    Item
    Algoritmos de inteligência computacional aplicados à otimização de sistemas de controle em acionamentos elétricos
    (Universidade Federal de Goiás, 2023-03-29) Santos, Guilherme Fernandes dos; Silva, Wander Gonçalves da; http://lattes.cnpq.br/4669127331497967; Cruz Junior, Gelson da; http://lattes.cnpq.br/4370555454162131; Cruz Junior, Gelson da; Silva, Wander Gonçalves da; Pickert, Volker; Oliveira, Marco Antonio Assfalk de; Cardoso, Alisson Assis
    This work presents the use of different computational intelligence methods applied to the tuning of a set of PI controllers for a DC motor drive with speed and position control. For position control, three closed control loops are used: armature current, speed and position. For speed control, only the armature current and speed loops are considered. In both cases, the outputs of the PI armature current and speed regulators are limited to the rated armature voltage and current, respectively. Then, it is possible to use higher gains for the controllers, what makes the system to respond faster. However, the windup phenomenon can arise. To avoid it, anti-windup circuits are also used and therefore, the system becomes non-linear. Because of this, an optimum tuning of the controllers may become a difficult task. In order to explore different possibilities, firstly, the speed and position control problems are formulated so that only one objective is minimised. Within this single-objective optimisation context, the PSO and SA algorithms are used to tune the controller parameters, them, the capability of each one is investigated when compared to each other. Multiobjective formulations are also explored to address three objectives simultaneously. In this part of the work, the multi-objective evolutionary algorithms NSGA-II and SPEA2 are used. All algorithms were implemented in MATLAB and the electric drive models were developed in the SIMULINK environment. Simulation results are presented showing that for the single-objective formulation, for both, the for speed and position control problems, the PSO algorithm outperformed the SA. For the multi-objective formulation, the SPEA2 algorithm presented better characteristics with respect to the Spread quality indicator in the only, when compared to NSGA-II. Furthermore, it’s shown to outperform the NSGA-II with respect to the Hypervolume indicator within the position control problem. A series of tests were carried out by varying the values of the main parameters setting for each algorithm. However, in most cases, no statistically significant advantage was observed. In general, the results presented demonstrate the ability of the algorithms to find optimal tuning for the controllers, either for the single-objective or the multiple and conflicting objectives problem.