Search

Introduction: The incorporation of an autonomous learning system in robotics will allow the resolution of a large number of problems. One is the autonomous march of the humanoid robots due to its complexity in the great number of variables regarding this process.Objective: Develop algorithms that generate autonomous paths in a humanoid robot with various degrees of freedom.Methodology: The study begins with the development of stochastic algorithms with few dimensions. Then, it will be extended to n-dimensional situations. Afterward, simulation tests will be carried out. And finally, the experimental tests are performed.Results: An algorithm was generated based on the physical model of the robot to create walking paths stochastically. A simulator that contemplates the kinematic constraints, including collisions, was implemented to verify the results. In addition, one hundred experimental tests were done. With these tests, the correct operation of the trajectories was verified.Conclusions: It was verified that it is possible to create a stochastic algorithm that mixes determinant and random rules to automatically generate paths in humanoid robots, hence, extending concepts generated in two-dimensional and three-dimensional spaces to n-dimensional articulated coordinates.

Cognitive radio is a paradigm that proposes maximizing the utilization of the usable radio-electric spectrum, allowing licensed users (PUs) and non-licensed users (SUs) to simultaneously coexist through the dynamic management and assignment of spectrum resources, by integrating the stages of spectrum sensing, decision, sharing and mobility. Spectrum decision is one of the most important stages, but its optimal operation depends on the characterization sub-stage, which is in charge of efficiently estimating time gaps in which a PU won’t make use of the assigned spectrum, so that it can be used in an opportunistic fashion by SUs. The design and implementation of an algorithm based on the Long Short-Term Memory (LSTM) recurrent neural network is proposed in order to increase the success percentage in the forecasting (presence/absence) of PUs in spectrum channels. The accuracy level exhibited in the results indicates LSTM increases the prediction percentage as compared to the Multilayer Perceptron Neural Network (MLPNN) and the Adaptative Neuro-Fuzzy Inference System (ANFIS) learning models, which means it could be implemented in cognitive networks with centralized physical topologies.

Objectives: This paper aims to solve a problem of existing research for Cognitive Radio networks Unlicensed (specifically for Wi-Fi with centralized network topology), which aims at the spectral making decision stage to efficiently characterize the behavior of users (called PUs) in order to identify spectral opportunities that can be used beneficially by other wireless applications or users. Methods/ Statistical Analysis: The methodology includes capturing the dynamics of channel usage by a user in the 2.4 GHz band, later to apply a pre-processing of the data and reach an estimate for the level of modeling (based training system) and prediction. Findings: The possible solution that was raised during the development of the research was to use a Support Vector Machine (SVM) to estimate the prediction by classification. Application/Improvements: The results suggest that the use of SVM does not correspond to the methodology of artificial intelligence accurately because the level of delivered estimate is not accurate.

In this work is carried out an investigation of the behavior of the volatile that emit the fruits contaminated in relation to those that are totally organic through an electronic nose, the aim is to achieve a qualitative classification into two classes, presence of pesticide and organic food, this which allows is to reduce the rates of diseases because of the excessive contamination of the food and create a new methodology through a low-cost electronic equipment that is a tool for small and medium-sized farmers.

In this paper, a new variable selection technique inspired in sequential forward selection but specifically designed to work with support vector machines is introduced. The usefulness of the variable selection coupled to support vector machines for solving classification and regression problems is assessed by analysing two different databases. The first database corresponds to different concentrations of vapours and vapour mixtures measured with a metal oxide gas-sensor e-nose and the second database corresponds to different Iberian hams measured with a mass-spectrometry based e-nose. Using a reduced set of important variables (i.e. reducing the dimensionality of input space by the variable selection procedure) results in support vector machines with better performance. For example, the success rate in ham classification (11-class problem) rises from 79.91% (when all the variables available are used) to 90.30% (when a reduced set of input variables is used). Furthermore, a quantitative analysis of ham samples with good accuracy is shown to be possible: when the variable selection process introduced is coupled to support vector machine regression models, the correlation coefficients of actual versus predicted humidity, water activity and salt in ham samples are 0.975, 0.972 and 0.943, respectively. This compares favourably with the correlation coefficients obtained when no variable selection is performed (0.937, 0.924 and 0.894).

We address the problem of variable selection within the pattern recognition engine of multisensor systems. This problem arises when there is high dimensionality in the data used for training (e.g. because a high number of sensors are used, or because many features are extracted from the response of each sensor, or both). Different variable selection techniques (including deterministic and stochastic methods) have been coupled with neural network-based classifiers. The usefulness of each technique implemented is benchmarked by evaluating its performance in terms of three objective parameters: the success rate in classification, the dimensionality of the final set of variables used for training and the time needed to complete the variable selection procedure. The database consisted of 96 measurements of ammonia, acetone and o-xylene vapours and their binary mixtures gathered using a metal oxide gas sensor array. Each measurement was described by 120 variables (12 sensors and 10 parameters each). A new strategy for variable selection, which is based on a two-step approach, is introduced that leads to the building of parsimonious classification models based on either the fuzzy ARTMAP or the probabilistic neural networks. For example, a 91.66% success rate in the simultaneous identification and quantification of the species and their mixtures was obtained using nine input variables only (out of the 120 available). This process of variable selection was conducted in two-step, i.e. a coarse selection based on a variance criterion followed by a simulated annealing process. This two-step variable selection took about 10 min to complete in a Pentium 4 PC platform.

Considerable attention has been given to strategies for variable selection in gas analysis using multisensor systems. Of the stochastic methods, genetic algorithms (GAs) have been found to be useful for variable selection, although they do not prevent irrelevant variables from being selected. Here we introduce the cascaded genetic algorithm, a new and simple approach that removes initially selected irrelevant variables. We coupled the method to the fuzzy ARTMAP classifier and used it to analyze single vapors and binary mixtures of three volatile organic compounds using a 12-element metal oxide gas sensor array. The method allowed to reduce the number of variables used to build the classifiers from 120 down to nine, which significantly increased the generalization ability of the fuzzy ARTMAP classifier. Success rates of 91.67 and 88.33% were reached in the simultaneous identification, quantification of single vapors, and in the identification of single vapors and their binary mixtures, respectively.

The morphologic and functional study of cardiovascular system is of vital importance because problems related to this system are one of the main causes of mortality in the world. It is important to mention that the left ventricle (VI) is the most susceptible to suffer severe damage, in diseases, such as arterial hypertension, mellitus diabetes or arteriosclerosis. This article presents a new methodology directed to segmentation of left ventricular contours, in angiographic images by using Generalized Hough Transform (TGH). It is important to obtain the ventricular edge, because analyzing processes of systole and diastole end, it is possible to calculate parameters of the cardiac functionality as the end-diastolic volume, end systolic volume, ejection fraction, cardiac output, Hyperkinéticos, Hypokinéticos segments, and normal; in this work we focus only on the removal of the same.

This paper deals with the evaluation of different bio-inspired algorithms techniques that allow reduce harmonic distortion (THD) in pulse width modulation (PWM) of power inverters. It is performed a comprehensive study which identifies and develops biologically inspired optimization algorithms based mainly on tabu search, genetic algorithms, particle swarm optimization and ant colony optimization. Bio-inspired optimization methods are used to find the best operational parameters of a PWM applied to a power inverter.

Industrial processes are characterized to be in open environments, with high uncertainty, unpredictability and nonlinear behavior. They have to be monitored and measured rigorously due to their behavior having a direct and serious impact on product quality, safety, productivity, pollution and finance. However, industrial processes have enormous volumes of complex and high dimensional data available, with poorly defined domains and redundant, noisy or inaccurate measures with unknown parameters. Therefore, using just relevant and informative variables will decrease the high dimensionality and will facilitate the use of techniques to find patterns in data to correctly identify the functional states of the process, improving the performance of monitoring and measuring tasks. In this paper, we address the problem of sensor selection in industrial processes, where a mathematical or structural model and the class labels are not available or suitable. We propose a wrapper feature selection approach based on clustering, to perform an accurate process dataset classification with minimal variables needed. The proposed method is applied on an intensification reactor, the ’open plate reactor (OPR)’, over thiosulfate and esterification reactions. Results are compared with previous work on the same datasets showing that fewer variables are needed to correctly identify all the functional states of the process.

High dimensionality is inherent to MS-based electronic nose applications where hundreds of variables per measurement (m/z fragments) - a significant number of them being highly correlated or noisy - are available. Feature selection is, therefore, an unavoidable pre-processing step if robust and parsimonious pattern classification models are to be developed. In this article, a new strategy for feature selection has been introduced and its good performance demonstrated using two MS e-nose databases. The feature selection is conducted in three steps. The first two steps are aimed at removing noisy, non-informative and highly collinear features (i.e., redundant), respectively. These two steps are computationally inexpensive and allow for dramatically reducing the number of variables (near 80% of initially available features are eliminated after the second step). The third step makes use of a stochastic variable selection method (simulated annealing) to further reduce the number of variables. For example, applying the method to an Iberian ham database has resulted in the number of features being reduced from 209 down to 14. Using the surviving m/z fragments, a fuzzy ARTMAP classifier was able to sort ham samples according to producer and quality (11-category classification) with a 97.24% success rate. The whole feature selection process runs in a few minutes in a Pentium IV PC platform. © 2006 Elsevier B.V. All rights reserved., This work was funded in part by CICYT under project no. TIC2003-06301, by the Thematic Network in Metabolism and Nutrition ref. C03/08 and by AECI under project no. 39/04/P/E

A new procedure for variable selection, which runs in two steps, is introduced. First, an unsupervised and very fast variable selection procedure is applied: a parameter that accounts for the correlation between the features available is computed and, only near 20% of initial variables (those that are less collinear) are retained for further selection. Then, a fine-tuning selection based on a deterministic method (stepwise) coupled to a simple probabilistic neural network is conducted on the variable subset that resulted from the first selection step. The method is demonstrated using a database consisting of vapors of ethanol, acetone and toluene and their binary mixtures (120 variables). Vapors can be simultaneously identified and quantified with a 95.83% success rate and the time needed for the whole process is about 5 minutes in a Pentium 4 PC platform. Being unsupervised, the fast variable selection method applies generally, even in aroma analysis problems where category discovery is an issue. This is illustrated by applying the method to mixture analysis using direct mass spectrometry

We introduce a new variable selection approach, which converges much faster to the optimal set of variables for a given application. The new procedure runs in two steps. First, a coarse and very fast variable selection procedure is applied: A figure of merit is defined and computed for every variable, a threshold value set and only the variables whose figure of merit is higher than the threshold are retained for further selection. Then, a fine-tuning selection based either on deterministic or stochastic methods is conducted on the variable subset that resulted from the first step. The method is demonstrated using a database consisting of vapors of ethanol, acetone and toluene and their binary mixtures (120 variables/measurement). Vapors can be simultaneously identified and quantified with a 92.7% success rate and the time needed for variable selection is reduced at least by a factor of 4.