nach oben

2017 | Buch

Kapitel lesen Erstes Kapitel lesen

Designing with Computational Intelligence

herausgegeben von: Nadia Nedjah, Heitor Silvério Lopes, Luiza de Macedo Mourelle

Verlag: Springer International Publishing

Buchreihe : Studies in Computational Intelligence

Enthalten in: Springer Professional "Wirtschaft+Technik" , Springer Professional "Technik" , Springer Professional "Wirtschaft"

Einloggen, um Zugang zu erhalten

Über dieses Buch

This book discusses a number of real-world applications of computational intelligence approaches. Using various examples, it demonstrates that computational intelligence has become a consolidated methodology for automatically creating new competitive solutions to complex real-world problems. It also presents a concise and efficient synthesis of different systems using computationally intelligent techniques.

Inhaltsverzeichnis

Frontmatter

Chapter 1. On Using Fuzzy Logic to Control a Simulated Hexacopter Carrying an Attached Pendulum

Abstract

Fuzzy logic is used in many applications from industrial process control to automotive applications, including consumers trend forecast, aircraft maneuvering control and others. Considering the increased interest in using of multi-rotor aircrafts (usually called drones) for many kinds of applications, it is important to study new methods to improve multi-rotor maneuverability while controlling its stability in a proper way. Controlling the flight of multi-rotors, specially those equipped six rotors, is not a trivial task. When considering the design of such a control systems, traditional approaches such as PD/PID are very difficult to design, in spite of being easily implementable.

Emanoel Koslosky, Marco A. Wehrmeister, João A. Fabro, André S. de Oliveira

Chapter 2. Monocular Pose Estimation for an Unmanned Aerial Vehicle Using Spectral Features

Abstract

Pose estimation of Unmanned Aerial Vehicles (UAV) using cameras is currently a very active research topic in computer and robotic vision, with special application in GPS-denied environments. However, the use of visual information for ego-motion estimation presents several difficulties, such as features search, data association (feature correlation), inhomogeneous features distribution in the image, etc.

Gastón Araguás, Claudio Paz, Gonzalo Perez Paina, Luis Canali

Chapter 3. Simultaneous Navigation and Mapping in an Autonomous Vehicle Based on Fuzzy Logic

Abstract

This research presents the navigation control and mapping of an autonomous car by fuzzy logic that enables automatic obstacle avoidance in unknown environments. The strategy is based on a map of the environment, which is created according to navigation, to plan the trajectories avoiding obstacles through the search algorithm A*. The proposed approach is evaluated in a virtual environment, where the autonomous car should move among different obstacles.

Álvaro Luiz Sordi Filho, Leonardo Presoto de Oliveira, André Schneider de Oliveira, João Alberto Fabro, Marco Aurélio Wehrmeister

Chapter 4. Fully Scalable Parallel Hardware for Wheeled Robot Navigation Using Fuzzy Control

Abstract

Process control is one of the many applications that took advantage of the fuzzy logic. Controllers are usually embedded into the controller device. This chapter aims at presenting the development of a reconfigurable efficient architecture for fuzzy controllers, suitable for embedding. The architecture is parameterizable so it allows the setup and configuration of the controller, so it can be used for various problem applications. An application of fuzzy controllers was implemented and its cost and performance have been evaluated.

Nadia Nedjah, Paulo Renato S. S. Sandres, Luiza de Macedo Mourelle

Chapter 5. Nonlinear Correction for an Energy Estimator Operating at Severe Pile-Up Conditions

Abstract

For systems operating at high event rates, the readout signal may be distorted by the presence of information from adjacent events. The signal superposition, or pile-up, degrades the efficiency of linear methods, which are typically used for signal parameter estimation. In many applications , the estimation task reduces to determine the amplitude of the incoming signal. In the context of high-energy calorimeters, which aim at measuring the energy of high-energy subproducts of interactions, the signal energy is measured by estimating the amplitude of the received digitized pulse. Modern particle colliders may operate at an event rate much higher than their calorimeter time response length and, as a result, the signal pile-up may be observed. This chapter describes how a computational intelligence approach can assist on the energy estimation performed by an optimal linear method. An artificial neural network is trained aiming at correcting for the nonlinearities introduced by the signal pile-up statistics. The efficiency of the various energy estimation methods is evaluated from simulation data under various signal pile-up scenarios.

Bernardo Sotto-Maior Peralva, Alessa Monay e Silva, Luciano Manhães de Andrade Filho, Augusto Santiago Cerqueira, José Manoel de Seixas

Chapter 6. Non-supervised Learning Applied to Analysis of Topological Metrics of Optical Networks

Abstract

Graphs can be used to model many real-world problems, such as social networks, telecommunication networks and biological structures. To aid the analysis of complex networks, several topological metrics and generational procedures have been proposed in the last years. This work offers a systematic method to analyse different backbone optical networks, based on a non-supervised algorithm for clustering and investigates the power of a recently proposed topological metrics, named \({I({\hat{\mathcal {F}}})}\). The metrics \({I({\hat{\mathcal {F}}})}\) and three others are applied to identify the best canonical model to represent real backbone optical networks. According to the obtained results, the clustering procedure allows to indicate \({I({\hat{\mathcal {F}}})}\) as the better metrics to explain the installed capacity for the analysed networks.

Danilo R. B. de Araújo, Joaquim F. Martins-Filho, Carmelo J. A. Bastos-Filho

Chapter 7. Mole Features Extraction for a Melanoma Recognition System

Abstract

The cancer is a painful disease that kill too many people. Skin cancer is among the most frequent types of tumors in the world, and melanoma is the most worrying type of skin cancer due to its high metastasis chances. Its global occurrence index is close to 133.000 people per year.

Henrique C. Siqueira, Bruno J. T. Fernandes

Chapter 8. Human–Machine Musical Composition in Real-Time Based on Emotions Through a Fuzzy Logic Approach

Abstract

In this chapter, a method for representing human emotions is proposed in the context of musical composition, which is used to artificially generate musical melodies through fuzzy logic. A real-time prototype system, for human–machine musical compositions, was also implemented to test this approach, using the emotional intentions captured from a human musician and later used to artificially compose and perform melodies accompanying a human artist while playing the chords.

Pedro Lucas, Efraín Astudillo, Enrique Peláez

Chapter 9. A Recursive Genetic Algorithm-Based Approach for Educational Timetabling Problems

Abstract

This chapter addresses the educational timetabling problem for multiple courses. This is a complex problem that basically involves a group of agents such as professors and lectures that must be weekly scheduled. The goal is to find solutions that satisfy the hard constraints and minimize the soft constraint violations. Moreover, universities often differ in terms of constraints and number of professors, courses, and resources involved, which increases the problem size and complexity. In this work, we propose a simple, scalable, and parameterized recursive approach to solve timetabling problems for multiple courses with genetic algorithms, which are efficient search methods used to achieve an optimal or near optimal solution.

Shara S. A. Alves, Saulo A. F. Oliveira, Ajalmar R. Rocha Neto

Chapter 10. Evolving Connection Weights of Artificial Neural Network Using a Multi-Objective Approach with Application to Class Prediction

Abstract

In Artificial Neural Network (ANN), the selection of connection weights is a key issue and Genetic and Evolution Strategies have been found to be promising algorithms to solve this important task. Motivated by that, this study investigates the applicability of using two novel Multi-Objective Evolutionary Algorithms (MOEA): Speed constrained Multi-Objective Particle Swarm Optimization (SMPSO) and Multi-Objective Differential Evolution Algorithm based on Decomposition with Dynamical Resource Allocation (MOEA/D-DE-DRA). ANNs are training to learn data classification using sensibility and specificity for different UCI databases. The results are compared using the Hypervolume as quality indicator and statistical test.

Andrei Strickler, Aurora Pozo

Chapter 11. Diversification Strategies in Evolutionary Algorithms: Application to the Scheduling of Power Network Outages

Abstract

The design of evolutionary algorithms that efficiently solve complex optimization problems can be considered a challenging puzzle. In complex and multimodal problems, premature convergence to a local optimum can compromise the search for better solutions. In this work, different strategies to avoid and/or fix premature convergence of evolutionary algorithms are proposed. High diversification level is maintained throughout the evolution process, so that an adequate trade-off between solution quality and computational cost is achieved. A metric that addresses diversification in evolutionary algorithms is employed. It is shown that this metric can be used to drive the search process conveniently. The proposed diversification strategies for evolutionary algorithms are tested in a real, complex, and epistatic scheduling problem concerned with the operation of power networks. Numerical results illustrate the application of the proposed strategies and respective impact on the quality and computational cost of solutions.

Rainer Zanghi, Julio Cesar Stacchini de Souza, Milton Brown Do Coutto Filho

Chapter 12. WBdetect: Particle Swarm Optimization for Segmenting Weld Beads in Radiographic Images

Abstract

The radiographic inspection of weld beads is important to ensure quality and safety in pipe networks. Visual fatigue, distractions, and the amount of radiographic images to be analyzed can be listed as main factors for human inspection errors. This chapter presents an approach for automatically segmenting weld beads in Double Wall Double Image (DWDI) X-ray photographs by combining two known methods in the literature: Particle Swarm Optimization (PSO) and Dynamic Time Warping (DTW). Vertical profiles of the weld beads are obtained from the windows’ coordinates encoded by particles and compared, via DTW, with a predefined model. Experiments are performed considering two phases: first, tests are carried out to set the default configuration, and second the configured system (named WBdetect) is evaluated, including a comparison with another approach. Promising results show that WBdetect converges, most of the time, to the window that allows a proper segmentation of the weld bead, outperforming the compared approach (the average accuracy achieved by WBdetect is 93.63 \(+-\)12.91, and 65.88 \(+-\)17.9 % for the other approach).

Rafael Miranda, Myriam Delgado, Tania Mezzadri, Ricardo Dutra da Silva, Marlon Vaz, Carla Marinho

Backmatter

Titel: Designing with Computational Intelligence
herausgegeben von: Nadia Nedjah
Heitor Silvério Lopes
Luiza de Macedo Mourelle
Verlag: Springer International Publishing
Electronic ISBN: 978-3-319-44735-3
Print ISBN: 978-3-319-44734-6
DOI: https://doi.org/10.1007/978-3-319-44735-3