Catálogo de publicaciones - libros

Surface roughness () control in High Speed Machining () demands reliable monitoring systems. A new data fusion model based on a multi-sensor system is developed. The model considers cutting parameters, cutting tool geometry, material properties and process variables. It can be used to predict the and  −process. The Response Surface Design methodology was used to minimize the number of experiments. Artificial neural networks were exploited as data fusion techniques. Early results represent the building blocks for a low cost supervisory control system that optimizes the in .

This paper describes a fault diagnosis system (FDS) for non-linear plants based on fuzzy logic. The proposed scheme, named VisualBlock-FIR, runs under the Simulink framework and enables early fault detection and identification. During fault detection, the FDS should recognize that the plant behavior is abnormal, and therefore, that the plant is not working properly. During fault identification, the FDS should conclude which type of failure has occurred. The enveloping and acceptability measures introduced in VisualBlock-FIR enhance the robustness of the overall process. The final part of this research shows how the proposed approach is used for tackling faults of the DAMADICS benchmark.

This paper proposes the use of a Recurrent Neural Network (RNN) for modeling a hydrocarbon degradation process carried out in a biopile system. The proposed RNN model represents a Kalman-like filter and it has seven inputs, five outputs and twelve neurons in the hidden layer, with global and local feedbacks. The learning algorithm is a modified version of the dynamic Backpropagation one. The obtained RNN model is simplified and used to design a Sliding Mode Control (SMC). The graphical simulation results of biopile system approximation, obtained via RNN model learning and the designed process SMC exhibited a good convergence, and precise system reference tracking.

In the last years Intelligent Tutoring Systems have been a very successful way for improving learning experience. Many issues must be addressed until this technology can be defined mature. One of the main problems within the Intelligent Tutoring Systems is the process of contents authoring: knowledge acquisition and manipulation process is a difficult task because it requires specialized skills on computer programming and knowledge engineering. In this paper we propose a mechanism based on first order data mining to partially automate the process of knowledge acquisition. The knowledge has to be used in the ITS during the tutoring process for personalized instruction. Such a mechanism can be applied in Constraint Based Tutor and in the Pseudo-Cognitive Tutor.

The development of intelligents teaching-learning systems depends, on one hand, of the pedagogical paradigms and, on the other hand, of the available technologies to implement these paradigms in computers. The field of the Intelligent Teaching-Learning Systems is characterized by the application of Artificial Intelligence techniques, to the development of the teaching-learning process assisted by computers, where the term "intelligent" is associated to the student’s aptitude to dynamically acclimatize to the teaching process by carrying out an individual learning. The case-based reasoning is an Artificial Intelligence technique that performs their reasoning process based on previously solved cases, stored in case-bases. In this article we propose a new case-based approach with foundations on fuzzy pattern recognition to help elaborate intelligents teaching-learning systems, using the -testor theory, based on a combination of typical testor theory with the fuzzy sets, assures the efficient access and retrieval of cases.

In this study, heuristic optimization methods which are genetic algorithm (GA), simulated annealing (SA) and adaptive simulated annealing genetic algorithm (ASAGA) are used for selecting questions from a question bank and generating a tets. The crossover and mutation operator of standard GA can not be directly usable for generating test, since integer-coded individuals have to be used and these operators produce duplicated genoms on individuals. In order to solve this problem, a mutation operation is proposed for preventing the duplications on crossovered individuals and also directing the search randomly to the new spaces. A database containing classified test questions is created together with predefined attributes for selecting questions. A particular test can be generated automatically, without active participation of the academician. The experiments and comparative analysis show that GA with proposed mutation operator is successful as nearly 100 percent and it produces results in noteworthy computational times.