Searcharxiv⌕ Search

arXiv subjects

E. G. Nepomuceno

Publications and source records attributed to E. G. Nepomuceno.

At least 19 recordsLinked to original sources
arXiv2024

Insightful Railway Track Evaluation: Leveraging NARX Feature Interpretation

The classification of time series is essential for extracting meaningful insights and aiding decision-making in engineering domains. Parametric modeling techniques like NARX are invaluable for comprehending intricate processes, such as environmental time series, owing to their easily interpretable and transparent structures. This article introduces a classification algorithm, Logistic-NARX Multinomial, which merges the NARX methodology with logistic regression. This approach not only produces interpretable models but also effectively tackles challenges associated with multiclass classification. Furthermore, this study introduces an innovative methodology tailored for the railway sector, offering a tool by employing NARX models to interpret the multitude of features derived from onboard sensors. This solution provides profound insights through feature importance analysis, enabling informed decision-making regarding safety and maintenance.

eess.SP↗
arXiv2021

Meta-Model Structure Selection: Building Polynomial NARX Model for Regression and Classification

This work presents a new meta-heuristic approach to select the structure of polynomial NARX models for regression and classification problems. The method takes into account the complexity of the model and the contribution of each term to build parsimonious models by proposing a new cost function formulation. The robustness of the new algorithm is tested on several simulated and experimental system with different nonlinear characteristics. The obtained results show that the proposed algorithm is capable of identifying the correct model, for cases where the proper model structure is known, and determine parsimonious models for experimental data even for those systems for which traditional and contemporary methods habitually fails. The new algorithm is validated over classical methods such as the FROLS and recent randomized approaches.

cs.LG↗
arXiv2021

Hybrid Method Based on NARX models and Machine Learning for Pattern Recognition

This work presents a novel technique that integrates the methodologies of machine learning and system identification to solve multiclass problems. Such an approach allows to extract and select sets of representative features with reduced dimensionality, as well as predicts categorical outputs. The efficiency of the method was tested by running case studies investigated in machine learning, obtaining better absolute results when compared with classical classification algorithms.

cs.LG↗
arXiv2020

Image encryption based on flexible computing of chaotic systems

The increase in data traffic on the internet has significantly increased the relevance of data and image encryption. Among the techniques most used in cryptography, chaotic systems have received great attention due to their easy implementation. However, it has recently been observed that these systems can lose their chaotic properties due to the finite precision of computers. In this work, we intend to investigate flexible computing tools, particularly interval analysis, to reduce this problem. We opted for the Lorenz System, as it is one of the few systems whose chaoticity is proven analytically. The results of this study, based on the correlation and entropy indexes, were superior to other studies published in the recent literature.

cs.CR↗
arXiv2019

Identificação de Sistemas Não Lineares Utilizando o Algoritmo Híbrido e Binário de Otimização por Enxame de Partículas e Busca Gravitacional

This work presents a new meta-heuristic approach to model structure selection of polynomial NARX models. In this respect, the technique penalizes the models based on the individual contribution of each regressor in representing the system. The new algorithm is tested on two experimental case studies: the identification of an electromechanical system and a eletric heater. The results are compared with Error Reduction Ratio and another meta-heuristic approach. The proposed method shows its advantages over compared methods in terms of the trade-off between prediction accuracy and model interpretability. The results are quantified and compared using the Mean Squared Error (MSE) indices.

eess.SY↗
arXiv2019

On the reliability of computational chaos-based cryptography for information exchange

This paper investigates the use of dynamical chaotic systems to encrypt and exchange images between different devices. Two devices were used to simulate the Cubic Map, having the same set of initial conditions, to generate an encryption key. Although both devices are floating-point compliant, the simulations, and consequently the encryption key, turned out to differ from one another. This indicates that many existing chaos-based encryption schemes are just special cases of computational arithmetic properties, in which some characteristics in the construction of the devices coincided. A method to mitigate such flaw was also presented.

cs.CR↗
arXiv2019

Some remarks on the performance of Matlab, Python and Octave in simulating dynamical systems

Matlab has been considered as a leader computational platform for many engineering fields. Well documented and reliable, Matlab presents as a great advantage its ability to increase the user productivity. However, Python and Octave are among some of the languages that have challenged Matlab. Octave and Python are well known examples of high-level scripting languages, with a great advantage of being open source software. The novelty of this paper is devoted to offer a comparison among these tree languages in the simulation of dynamical systems. We have applied the lower bound error to estimate the error of simulation. The comparison was performed with the chaotic systems Duffing-Ueda oscillator and the Chua's circuit, both identified with polynomial NARMAX. Octave presents the best reliable outcome. Nevertheless, Matlab needs the lowest time to undertake the same activity. Python has presented the worse result for the stop simulation criterion.

cs.MS↗
arXiv2019

Simplification of the digital representation of the tent map through biased fixed point

Chaotic systems have been investigated in several areas of engineering. In control theory, such systems have instigated the emergence of new techniques as well, have been used as a source of noise generation. The application of chaotic systems as pseudo-random numbers has also been widely employed in cryptography. One of the central aspects of these applications in high performance situations, such as those involving a large amount of data (Big Data), is the response of these systems in a short period of time. Despite the great advances in the design of chaotic systems in analog circuits, it is perceived less attention in the optimized design of these systems in the digital domain. In this work, the polarized fixed point representation is applied to reduce the number of digital elements. Using this approach, it was possible to significantly reduce the number of logic gates in the subtraction operation. When compared to other works in the literature, it has been viable to reduce by 50 \% the number of elements per bit of the digital representation of the tent map. The chaoticity was evidenced with the calculation of the Lyapunov exponent. Histogram, entropy and autocorrelation tests were used satisfactorily to evaluate the randomness of the represented system.

eess.SP↗
arXiv2019

Detection of Muscle Fatigue Using Variable Bit Flow Modulation and Cross Correlation of Electromyographic Signals

The surface electromyography (sEMG) analysis can provide information on muscle fatigue status by estimation of muscle fibre conduction velocity (MFCV), a measure of the travelling speed of motor unit action potentials in muscle tissue. This paper proposes a technique for MFCV estimation using cross-correlation methods and variable bitstream modulation. The technique displays an estimate based on a set of data generated by the gain variation modulation, providing an average estimate of the MFCV. The observed trend of MFCV decrease correlates with the fatigue state of the observed muscle. Finally, the values found were compared with information from the literature, validating the method and showing the advantages of using variable modulation.

eess.SP↗
arXiv2019

Simulation Reproducibility of a Chaotic Circuit

An evergreen scientific feature is the ability for scientific works to be reproduced. This feature allows researchers to understand, enhance, or even question works that have been developed by other scientists. In control theory the importance of modeling and simulation of systems is widely recognized. Despite this recognition, less attention is paid to the effects of finite precision of computers on the simulation reproducibility of nonlinear dynamic systems. In this work, a case study of reproducibility is presented in the simulation of a chaotic Jerk circuit, using the software LtSpice. In order to do so, we performed simulations of the circuit in the same version of the software on different computers, in order to collect the data and compare them with experimental results. The comparison was made with the NRMSE (Normalized Root Mean Square Error), in order to identify the computer with the highest prediction horizon. Tests performed in 4 different configurations showed the difficulties of simulation reproducibility in LtSpice. The methodology developed was efficient in identifying the computer with better performance, which allows applying it to other cases in the literature.

cs.OH↗
arXiv2019

Computation of the largest Lyapunov exponent using recursive estimation with variable factor

Chaotic systems have been investigated in the most diverse areas. One of the first steps in chaotic system research is the detection of chaos. The largest Lyapunov exponent (LLE) is one of the most widely used techniques for this purpose. Recently, techniques for calculating LLE have been developed taking into account the error due to the finite precision of computers. Recursive methods were employed to improve such algorithms. However, this method uniformly weighed the data used to calculate the LLE. This paper investigates the different weighing of the data based on the hypothesis that the initial data have a higher precision than the last data processed by the algorithm, since the process is subject to error accumulation. In five tested systems, the proposed method obtained more accurate results in three. However, there was an increase in the variance of the result obtained for two of the evaluated systems.

math.NA↗
arXiv2019

A Survey of the Individual-Based Model applied in Biomedical and Epidemiology

Individual-based model (IBM) has been used to simulate and to design control strategies for dynamic systems that are subject to stochasticity and heterogeneity, such as infectious diseases. In the IBM, an individual is represented by a set of specific characteristics that may change dynamically over time. This feature allows a more realistic analysis of the spread of an epidemic. This paper presents a literature survey of IBM applied to biomedical and epidemiology research. The main goal is to present existing techniques, advantages and future perspectives in the development of the model. We evaluated 89 articles, which mostly analyze interventions aimed at endemic infections. In addition to the review, an overview of IBM is presented as an alternative to complement or replace compartmental models, such as the SIR (Susceptible-Infected-Recovered) model. Numerical simulations also illustrate the capabilities of IBM, as well as some limitations regarding the effects of discretization. We show that similar side-effects of discretization scheme for compartmental models may also occur in IBM, which requires careful attention.

q-bio.PE↗
arXiv2018

Image Encryption Algorithm Using Natural Interval Extensions

It is known that chaotic systems have widely been used in cryptography. Generally, floating point simulations are used to generate pseudo-random sequence of numbers. Although, it is possible to find some works on the degradation of chaotic systems due to finite precision of digital computers, little attention has been paid to exploit this limitation to formulate efficient process for image encode. This article proposes a novel image encryption method using natural interval extensions. The sequence of arithmetic operations is different in each natural interval extension. This is what we need to produce two different sequences; the difference between these sequences is used to generate the lower bound error, which has been shown to present satisfactory pseudo-random properties. The approach has been successfully tested using the Chua's circuit as the chaotic system. The secret key has presented good properties for encrypting the Lena image.

cs.CR↗
arXiv2018

Decimation analysis in the signal processing of current to detect broken bars in induction machine

This paper presents a study on the reduction of the sampling frequency of the current signals of an induction motor, the reductions are performed by means time-decimation technique for digital signal processing. We have used the Fast Fourier Transform to obtain the fault signal spectrum of broken bars. The results have shown how the decimation technique significantly reduces the number of operations and the time required to calculate the Fast Fourier Transform without loss of information. This approach provides a better performance of embedded systems for fault diagnosis based on characteristic of amplitude signal modulation.

eess.SP↗
arXiv2018

Interval Simulation of Narmax Models Based on Computer Arithmetic

System identification is an important area of science, which aims to describe the characteristics of the system, representing them by mathematical models. Since many of these models can be seen as recursive functions, it is extremely important to control the errors in these functions, because small errors introduced in each computational step can grow exponentially due to the sensitivity to initial conditions present in this type of functions. One of the ways to control rounding and truncation errors is through interval arithmetic, since it is not possible to represent all numbers in the computer because of the finite representation in them. Thus, in arithmetic interval a number is represented by an interval in which the true number is within that interval. In this manuscript we developed an algorithm that performs the operations of interval arithmetic using basic functions. We have compared compared our results with the Matlab-toolbox Intlab. Numerical experiments have shown that our method is superior producing narrower intervals.

eess.SP↗
arXiv2017

The Lower Bound Error for polynomial NARMAX using an Arbitrary Number of Natural Interval Extensions

The polynomial NARMAX (Nonlinear AutoRegressive Moving Average model with eXogenous input) is a model that represents the dynamics of physical systems. This polynomial contains information from the past of the inputs and outputs of the process, that is, it is a recursive model. In digital computers this generates the propagation of the rounding error. Our procedure is based on the estimation of the maximum value of the lower bound error considering an arbitrary number of pseudo-orbits produced from different natural interval extensions, and a posterior Lyapunov exponent calculation. We applied successfully our technique for two identified models of the systems: sine map and Duffing-Ueda oscillator

eess.SP↗
arXiv2016

Simulation of Dynamical Systems with Interval Analysis: A case study of RLC Circuit

Differences between computer simulation of dynamical systems and laboratory experiments are common in teaching and research in engineering. Normally, numerical inaccuracy and the non-ideal behaviour of the devices involved in the experiment are the most common explanations. With the application of interval analysis, it is possible to incorporate the numerical and parametric uncertainties in the simulation, allowing a better understanding of the play between simulation and experiment. This article presents a case study in which an step input is applied to an RLC circuit. Using the toolbox Intlab for Matlab, it was possible to present a computer simulation with the range that encompasses the experimental results . Comparison of simulation with experimental data show the success of the technique and indicates a potential content to be delivered to undergraduate engineering courses.

math.NA↗