Catálogo de publicaciones - libros

A self-organizing medical diagnosis system, mirroring swarm intelligence to structure knowledge in holonic patterns is presented. The system sets up on an alliance of medical experts – realized by agents – that stigmergically self-organize in order to provide a viable medical diagnosis. Starting point is always a flat set of autonomous agents that spontaneously and temporarily organize on a case-based basis into a holarchy (without hierarchical control flow). Despite their sophisticated task the proposed agents, like ants in an ant colony, exhibit a comparatively simple architecture built on reactive behavior. The real power of the system stems from the fact that a large number of those simple agents collaborate in order to come to a reliable diagnosis.

We conducted data mining method (association rule analysis) to elucidate the relationship between 6 lifestyles (overweight, drinking, smoking, meals, physical exercise, sleeping time, and meals), 5 family medical histories (hypertension, diabetes, cardiovascular disease, cerebrovascular disease, and liver disease), and 6 medical abnormalities (high blood pressure, hyperchoresterolemia, hypertrigriceridemia, high blood sugar, hyperuricemia, and liver dysfunction) in examination data using the medical examination data of 7 years, obtained from 5,350 male employees in the age group of 40-49 years. We found that number of combinations derived from data mining (association rule method) was greater than that derived from conventional method (logistic regression analysis). Moreover, values of both “confidence” and “odds ratio” derived from association rule were greater than that derived from logistic regression. We found that “the association rule method” was more and useful to elucidate effective combinations of risk factors in terms of lifestyle diseases.

The ability of a moving-mass control system to control a spinning vehicle using two internal moving mass actuators is investigated. The nonlinear equations of motion are provided, and the influence to the system of moving masses’ motion with respect to the vehicle’s shell is described. For the self-learning capacity of the neural networks and the optimum ability of the genetic algorithm, the hybrid trajectory PID control scheme based on the neural networks and genetic algorithm is produced to improve the dynamic qualities and the adaptive capacity of the system. A nonlinear simulation of a typical mission profile demonstrates the ability of the controller to effectively control the vehicle’s trajectory.

Fitting brightness profiles of galaxies in one dimension is frequently done because it suffices for some applications and is simple to implement, but many studies now resort to two-dimensional fitting, because many well-resolved, nearby galaxies are often poorly fitted by standard one-dimensional models. For the fitting we use a model based on de Vaucoleurs and exponential functions that is represented as a set of concentric generalized ellipses that fit the brightness profile of the image. In the end, we have an artificial image that represents the light distribution in the real image, then we make a comparison between such artificial image and the original to measure how close the model is to the real image. The problem can be seen as an optimization problem because we need to minimize the difference between the original optical image and the model, following a normalized Euclidean distance. In this work we present a solution to such problem from a point of view of optimization using a hybrid algorithm, based on the combination of Evolution Strategies and the Quasi-Newton method. Results presented here show that the hybrid algorithm is very well suited to solve the problem, because it can find the solutions in almost all the cases and with a relatively low cost.

The clustering problem under the criterion of minimum sum of squares clustering is a nonconvex program which possesses many locally optimal values, resulting that its solution often falls into these traps. In this paper, a hybrid tabu search based clustering algorithm called KT-Clustering is developed to explore the proper clustering of data sets. Based on the framework of tabu search, KT-Clustering gathers the optimization property of tabu search and the local search capability of K-means algorithm together. Moreover, mutation operation is adopted to establish the neighborhood of KT-Clustering. Its superiority over K-means algorithm, a genetic clustering algorithm and another tabu search based clustering algorithm is extensively demonstrated for experimental data sets.

To cope with resource constraints in multitasking control systems, feedback scheduling is emerging as an important technique for integrating control and scheduling. The ability of neural networks (NNs) as good and robust nonlinear function approximators, reducing the computation time as compared against algorithmic solutions, suggests applying them to the feedback scheduling problem. A novel, simple and intelligent feedback scheduler is presented using a feedforward NN. The algorithmic optimizer is utilized as a teacher to generate data samples for NN training. The role of the NN based feedback scheduler is to provide a good approximation to the optimal solution and online adjust the sampling period of each control task so that the overall system performance is maximized in the face of workload variations. The performance of the NN approach is evaluated through co-simulations of the scheduler, controllers and process dynamics.

We propose a “Hardware-In-the-Loop Simulation” (HILS) and “Rapid Control Prototyping” (RCP) based inter-working scheme for computational evaluation of manipulators and trajectory controller. The objective of the scheme is to minimize the development time in controller design and reduce the efforts in body modeling analysis, especially in the field of robotics. We have derived the analytical model of two manipulators for HILS and implemented an output controller for RCP; then two functional blocks are integrated to verify and validate the usefulness and possibility of HILS and RCP based inter-working (HRI) scheme. Experimental results show that the proposed HRI scheme is an effective means to apply for modeling the two elbow manipulators and implementing the selected controller for the manipulator applications in an early stage of the development cycle. Furthermore, HRI scheme has shown to be particularly beneficial for body dynamics analysis and manipulator control tasks.

Modeling of nonlinear static system using neural network based intelligent technology is presented in this paper. The architecture of the intelligent system is combined neural network with polynomial neural network. The composite architecture is designed to get a heuristic approximation method for nonlinear static system modeling. Owing to the approximation capabilities, neural networks have been widely utilized to process modeling, whereas the polynomial neural network is an analysis technique for identifying nonlinear relationships between inputs and outputs of the target system. So the hybrid architecture can harmonize the advantages of the each modeling methodology. Simulation results of the intelligent technology will be shown efficient and good performance.

This paper presents a technique for an intelligent robot to adaptively behave in unforeseen and dynamic circumstances. Since the traditional methods utilized the relatively reliable information about the environment to control intelligent robots, they were robust but could not behave adaptively in complex and dynamic world. On the contrary, behavior-based approach is suitable for generating autonomous behaviors in the environment, but it still lacks of the capabilities to infer dynamic situations for high-level behaviors. This paper proposes a 2-layer control architecture to generate adaptive behaviors, which perceive and avoid dynamic moving obstacles as well as static obstacles. The first level is to generate reflexive and autonomous behaviors with the behavior network, and the second level is to infer dynamic situation of mobile robots with Bayesian network. Experimental results with various situations have shown that the robot reaches the goal points while avoiding static or moving obstacles with the proposed architecture.

This paper deals with the GA-PSO (Genetic Algorithm-Particle Swarm Optimization) based vector control for loss minimization operation of induction motor. It is estimated that more than around 50% of the world electric energy generated is consumed by electric machines such as induction motor, DC motor. So, improving efficiency in electric drives is important and control strategy for minimum energy loss is needed as one of optimal operation strategy. In this paper, vector control approach is suggested for an optimal operation of induction motor using GA-PSO tuning method through simulation.