Showing 1–11 of 11 results for author: Martins, S A M

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

Authors: W. R. Lacerda Junior, S. A. M. Martins, E. G. Nepomuceno

Abstract: 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… ▽ More 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. △ Less

Submitted 20 September, 2021; originally announced September 2021.

Modelagem de Sistemas Audiométricos Usando Técnicas de Computação Flexível

Authors: Erick Schultz S. A. Caetano, Denise Fonseca Resende, Samir Angelo Milani Martins, Erivelton Geraldo Nepomuceno, Leonardo Bonato Felix

Abstract: In systems identification, the studied phenomena are accompanied by uncertainties, whether arising from measurement data or computational calculations. Interval data provides a valuable way to represent available information on complex problems where uncertainty, inaccuracy, or variability must be taken into account. The present work aims to determine interval parameters for a model considering da… ▽ More In systems identification, the studied phenomena are accompanied by uncertainties, whether arising from measurement data or computational calculations. Interval data provides a valuable way to represent available information on complex problems where uncertainty, inaccuracy, or variability must be taken into account. The present work aims to determine interval parameters for a model considering data measurement uncertainties using the MQ estimator and neural networks. The main objective of this work is to apply this technique in audiometric systems, particularly in the automatic detection of auditory responses taking into consideration the flexible computing concepts that offer solutions tolerant to subjectivity, inaccuracy or uncertainty. It was possible to obtain a model with interval parameters that allow an infinite set of parameters to be evaluated as a limited range. The model was validated using two methods, one with prediction of 2 steps ahead, using the MQ estimator and extended to another representation as neural networks 5 steps delay. It has been shown that interval parameters generate interval results that contain most validation data improving system reliability. △ Less

Submitted 20 November, 2019; originally announced November 2019.

Comments: 16 pages, in Portuguese, 9 figures, To appear in the proceedings of II Latin American Workshop on Computational Neuroscience, São João del-Rei, MG - Brazil - September, 18-20, 2019

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

Authors: W. R. Lacerda Junior, S. A. M. Martins, E. G. Nepomuceno

Abstract: 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 w… ▽ More 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. △ Less

Submitted 12 November, 2019; originally announced November 2019.

Comments: in Portuguese, SBAI 2019 - Simpósio Brasileiro de Automação Inteligente - Ouro Preto. 6 pages

Interval Simulation of Narmax Models Based on Computer Arithmetic

Authors: P. F. S. Guedes, M. L. C. Peixoto, O. A. R. O. Freitas, A. M. Barbosa, S. A. M. Martins, E. G. Nepomuceno

Abstract: 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 initia… ▽ More 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. △ Less

Submitted 25 July, 2018; originally announced July 2018.

Comments: CBA 2018 - XXII Congresso Brasileiro de Automatica, Joao Pessoa, p. 1-6

Synchronization on the accuracy of chaotic oscillators simulations

Authors: Gabriel H. A. Silva, Igor C. Silva, Wilson R. L. Junior, Samir A. M. Martins, Marcio F. S. Barroso, Erivelton G. Nepomuceno

Abstract: Numerical problems are considered on general synchronization of chaotic oscillators, through the evaluation of the Lower Bound Error index on two case studies: a Lorenz system unidirectionally coupled to a Duffing system and a Duffing system unidirectionally coupled to a Rossler system. It was possible to observe, in each case, that the behavior of the slave's LBE curve tends to follow the behavio… ▽ More Numerical problems are considered on general synchronization of chaotic oscillators, through the evaluation of the Lower Bound Error index on two case studies: a Lorenz system unidirectionally coupled to a Duffing system and a Duffing system unidirectionally coupled to a Rossler system. It was possible to observe, in each case, that the behavior of the slave's LBE curve tends to follow the behavior of the master's as the value of the coupling constant is increased up to a certain value, and thus, that synchronization can affect numerical calculations. △ Less

Submitted 4 December, 2017; originally announced December 2017.

Comments: DINCON 2017 - Conferencia Brasileira de Dinamica, Controle e Aplicacoes - Sao Jose do Rio Preto - Brazil. 8 pages

Note on improvement precision of recursive function simulation in floating point standard

Authors: Melanie R. Silva, Erivelton G. Nepomuceno, Samir A. M. Martins

Abstract: An improvement on precision of recursive function simulation in IEEE floating point standard is presented. It is shown that the average of rounding towards negative infinite and rounding towards positive infinite yields a better result than the usual standard rounding to the nearest in the simulation of recursive functions. In general, the method improves one digit of precision and it has also bee… ▽ More An improvement on precision of recursive function simulation in IEEE floating point standard is presented. It is shown that the average of rounding towards negative infinite and rounding towards positive infinite yields a better result than the usual standard rounding to the nearest in the simulation of recursive functions. In general, the method improves one digit of precision and it has also been useful to avoid divergence from a correct stationary regime in the logistic map. Numerical studies are presented to illustrate the method. △ Less

Submitted 1 December, 2017; originally announced December 2017.

Comments: DINCON 2017 - Conferencia Brasileira de Dinamica, Controle e Aplicacoes - Sao Jose do Rio Preto - Brazil. 8 pages

Chaos suppression of Lorenz Systems by means on the average of rounding modes

Authors: Melanie R. Silva, Erivelton G. Nepomuceno, Samir A. M. Martins

Abstract: This work deals with chaos suppression based on average of the rounded modes to negative and positive infinite. The present procedure acts to reduce the rounding errors. It was observed that when the method proposed in this paper is applied to the chaotic Lorenz's system, it exhibits a periodic behaviour, characterized by a limit cycle and negative largest Lyapunov exponent. We tested our approach… ▽ More This work deals with chaos suppression based on average of the rounded modes to negative and positive infinite. The present procedure acts to reduce the rounding errors. It was observed that when the method proposed in this paper is applied to the chaotic Lorenz's system, it exhibits a periodic behaviour, characterized by a limit cycle and negative largest Lyapunov exponent. We tested our approach using three discretization schemes based on Runge-Kutta method △ Less

Submitted 25 November, 2017; originally announced November 2017.

Comments: DINCON 2017 - Conferencia Brasileira de Dinamica, Controle e Aplicacoes - Sao Jose do Rio Preto - Brazil. 8 pages

Incorporating Numerical Uncertainties for Validation of Nonlinear Models

Authors: Igor C. Silva, Gabriel H. A. Silva, Samir A. M. Martins, Erivelton G. Nepomuceno

Abstract: This paper proposes a model validation method that incorporates error due to numerical procedures. Two identified models for Sine Map and Duffing-Ueda Circuit systems have been investigated. The indexes RMSE and MAPE have been applied. We have shown that after some few iterates, it is possible to notice some significative difference between index provided in the literature. This difference has bee… ▽ More This paper proposes a model validation method that incorporates error due to numerical procedures. Two identified models for Sine Map and Duffing-Ueda Circuit systems have been investigated. The indexes RMSE and MAPE have been applied. We have shown that after some few iterates, it is possible to notice some significative difference between index provided in the literature. This difference has been computed in around 34%. △ Less

Submitted 21 November, 2017; originally announced November 2017.

Comments: DINCON 2017 - Conferencia Brasileira de Dinamica, Controle e Aplicacoes - Sao Jose do Rio Preto - Brazil. 7 pages. In Portuguese

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

Authors: Priscila F. S. Guedes, M. L. C. Peixoto, A. M. Barbosa, S. A. M. Martins, E. G. Nepomuceno

Abstract: 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 th… ▽ More 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 △ Less

Submitted 20 November, 2017; originally announced November 2017.

Comments: DINCON 2017 - Conferencia Brasileira de Dinamica, Controle e Aplicacoes. Sao Jose do Rio Preto. Brazil. 7 pages

Analysis of RLD Circuit by means of Phase Space with a non-ideal diode model

Authors: Marcia L. C. Peixoto, Erivelton G. Nepomuceno, Samir A. M. Martins

Abstract: In this work the non-linear and chaotic behaviour for the RLD circuit is analysed, which is modelled by state space and numerically simulated considering the diode model an association between nonlinear resistances and capacitances. It has been possible to characterize the dynamics of the circuit through period doubles and chaotic behaviour. The Lyapunov exponent is calculated to show the chaotic… ▽ More In this work the non-linear and chaotic behaviour for the RLD circuit is analysed, which is modelled by state space and numerically simulated considering the diode model an association between nonlinear resistances and capacitances. It has been possible to characterize the dynamics of the circuit through period doubles and chaotic behaviour. The Lyapunov exponent is calculated to show the chaotic behaviour of this circuit. Simulated results are in good agreement with experimental data. △ Less

Submitted 19 November, 2017; originally announced November 2017.

Comments: DINCON 2017 - Conferencia Brasileira de Dinamica, Controle e Aplicacoes - Sao Jose do Rio Preto - Brazil. 7 pages. In Portuguese

Identification of Dynamic Systems with Interval Arithmetic

Authors: Márcia L. C. Peixoto, Marco T. R. Matos, Wilson R. Lacerda Júnior, Samir A. M. Martins, Erivelton G. Nepomuceno

Abstract: This paper aims to identify three electrical systems: a series RLC circuit, a motor/generator coupled system, and the Duffing-Ueda oscillator. In order to obtain the system's models was used the error reduction ratio and the Akaike information criterion. Our approach to handle the numerical errors was the interval arithmetic by means of the resolution of the least squares estimation. The routines… ▽ More This paper aims to identify three electrical systems: a series RLC circuit, a motor/generator coupled system, and the Duffing-Ueda oscillator. In order to obtain the system's models was used the error reduction ratio and the Akaike information criterion. Our approach to handle the numerical errors was the interval arithmetic by means of the resolution of the least squares estimation. The routines was implemented in Intlab, a Matlab toolbox devoted to arithmetic interval. Finally, the interval RMSE was calculated to verify the quality of the obtained models. The applied methodology was satisfactory, since the obtained intervals encompass the system's data and allow to demonstrate how the numerical errors affect the answers. △ Less

Submitted 8 August, 2017; originally announced August 2017.

Comments: SBAI 2017 - XIII Simpósio Brasileiro de Automação Inteligente, Porto Alegre, p. 1-6, In Portuguese. arXiv admin note: text overlap with arXiv:1612.02674