Identification of a cylindrical robot using recurrent neural networks

Carlos Román Mariaca Gaspar; Juan Eduardo Velázquez-Velázquez; Julio César Tovar Rodríguez

doi:10.1007/978-3-319-09858-6_36

Identification of a cylindrical robot using recurrent neural networks

Carlos Román Mariaca Gaspar, Juan Eduardo Velázquez-Velázquez, Julio César Tovar Rodríguez

Unidad Profesional Interdisciplinaria de Ingeniería, Campus Hidalgo (UPIIH)

Producción científica: Capítulo del libro/informe/acta de congreso › Contribución a la conferencia › revisión exhaustiva

1 Cita (Scopus)

Resumen

Neural identification techniques are very useful for the problem of unknown dynamics and uncertainties during the development of a model that accurately represents the behaviour of a robot. In this paper we use the model of a Recurrent Trainable Neural Network (RTNN) for modelling a cylindrical robot. The RTNN proposal is a multilayer network local feedback into the single hidden layer, to approach the robot dynamics. The learning algorithm for this topology is the Backpropagation (BP) dynamic. The simulation results of the approximation obtained through RTNN showed a good convergence and accurate tracking.

Idioma original	Inglés
Título de la publicación alojada	Multibody Mechatronic Systems - Proceedings of the MUSME Conference, 2014
Editores	Eusebio Eduardo Hernández Martinez, Marco Ceccarelli
Editorial	Kluwer Academic Publishers
Páginas	381-389
Número de páginas	9
ISBN (versión digital)	9783319098579
DOI	https://doi.org/10.1007/978-3-319-09858-6_36
Estado	Publicada - 2015
Evento	5th International Symposium on Multibody Systems and Mechatronics, MUSME 2014 - Huatulco, México Duración: 21 oct. 2014 → 24 oct. 2014

Serie de la publicación

Nombre	Mechanisms and Machine Science
Volumen	25
ISSN (versión impresa)	2211-0984
ISSN (versión digital)	2211-0992

Conferencia

Conferencia	5th International Symposium on Multibody Systems and Mechatronics, MUSME 2014
País/Territorio	México
Ciudad	Huatulco
Período	21/10/14 → 24/10/14

Acceder al documento

10.1007/978-3-319-09858-6_36

Otros archivos y enlaces

Enlace a la publicación en Scopus

Citar esto

Gaspar, C. R. M., Velázquez-Velázquez, J. E., & Rodríguez, J. C. T. (2015). Identification of a cylindrical robot using recurrent neural networks. En E. E. H. Martinez, & M. Ceccarelli (Eds.), Multibody Mechatronic Systems - Proceedings of the MUSME Conference, 2014 (pp. 381-389). (Mechanisms and Machine Science; Vol. 25). Kluwer Academic Publishers. https://doi.org/10.1007/978-3-319-09858-6_36

Gaspar, Carlos Román Mariaca ; Velázquez-Velázquez, Juan Eduardo ; Rodríguez, Julio César Tovar. / Identification of a cylindrical robot using recurrent neural networks. Multibody Mechatronic Systems - Proceedings of the MUSME Conference, 2014. editor / Eusebio Eduardo Hernández Martinez ; Marco Ceccarelli. Kluwer Academic Publishers, 2015. pp. 381-389 (Mechanisms and Machine Science).

@inproceedings{09f3ce7add144bd9a1840667456d997f,

title = "Identification of a cylindrical robot using recurrent neural networks",

abstract = "Neural identification techniques are very useful for the problem of unknown dynamics and uncertainties during the development of a model that accurately represents the behaviour of a robot. In this paper we use the model of a Recurrent Trainable Neural Network (RTNN) for modelling a cylindrical robot. The RTNN proposal is a multilayer network local feedback into the single hidden layer, to approach the robot dynamics. The learning algorithm for this topology is the Backpropagation (BP) dynamic. The simulation results of the approximation obtained through RTNN showed a good convergence and accurate tracking.",

keywords = "Identification systems, Manipulator robot, Neural networks",

author = "Gaspar, {Carlos Rom{\'a}n Mariaca} and Vel{\'a}zquez-Vel{\'a}zquez, {Juan Eduardo} and Rodr{\'i}guez, {Julio C{\'e}sar Tovar}",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing Switzerland 2015.; 5th International Symposium on Multibody Systems and Mechatronics, MUSME 2014 ; Conference date: 21-10-2014 Through 24-10-2014",

year = "2015",

doi = "10.1007/978-3-319-09858-6_36",

language = "Ingl{\'e}s",

series = "Mechanisms and Machine Science",

publisher = "Kluwer Academic Publishers",

pages = "381--389",

editor = "Martinez, {Eusebio Eduardo Hern{\'a}ndez} and Marco Ceccarelli",

booktitle = "Multibody Mechatronic Systems - Proceedings of the MUSME Conference, 2014",

address = "Pa{\'i}ses Bajos",

}

Gaspar, CRM, Velázquez-Velázquez, JE & Rodríguez, JCT 2015, Identification of a cylindrical robot using recurrent neural networks. En EEH Martinez & M Ceccarelli (eds.), Multibody Mechatronic Systems - Proceedings of the MUSME Conference, 2014. Mechanisms and Machine Science, vol. 25, Kluwer Academic Publishers, pp. 381-389, 5th International Symposium on Multibody Systems and Mechatronics, MUSME 2014, Huatulco, México, 21/10/14. https://doi.org/10.1007/978-3-319-09858-6_36

Identification of a cylindrical robot using recurrent neural networks. / Gaspar, Carlos Román Mariaca; Velázquez-Velázquez, Juan Eduardo; Rodríguez, Julio César Tovar.
Multibody Mechatronic Systems - Proceedings of the MUSME Conference, 2014. ed. / Eusebio Eduardo Hernández Martinez; Marco Ceccarelli. Kluwer Academic Publishers, 2015. p. 381-389 (Mechanisms and Machine Science; Vol. 25).

Producción científica: Capítulo del libro/informe/acta de congreso › Contribución a la conferencia › revisión exhaustiva

TY - GEN

T1 - Identification of a cylindrical robot using recurrent neural networks

AU - Gaspar, Carlos Román Mariaca

AU - Velázquez-Velázquez, Juan Eduardo

AU - Rodríguez, Julio César Tovar

N1 - Publisher Copyright: © Springer International Publishing Switzerland 2015.

PY - 2015

Y1 - 2015

N2 - Neural identification techniques are very useful for the problem of unknown dynamics and uncertainties during the development of a model that accurately represents the behaviour of a robot. In this paper we use the model of a Recurrent Trainable Neural Network (RTNN) for modelling a cylindrical robot. The RTNN proposal is a multilayer network local feedback into the single hidden layer, to approach the robot dynamics. The learning algorithm for this topology is the Backpropagation (BP) dynamic. The simulation results of the approximation obtained through RTNN showed a good convergence and accurate tracking.

AB - Neural identification techniques are very useful for the problem of unknown dynamics and uncertainties during the development of a model that accurately represents the behaviour of a robot. In this paper we use the model of a Recurrent Trainable Neural Network (RTNN) for modelling a cylindrical robot. The RTNN proposal is a multilayer network local feedback into the single hidden layer, to approach the robot dynamics. The learning algorithm for this topology is the Backpropagation (BP) dynamic. The simulation results of the approximation obtained through RTNN showed a good convergence and accurate tracking.

KW - Identification systems

KW - Manipulator robot

KW - Neural networks

UR - http://www.scopus.com/inward/record.url?scp=84928238742&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-09858-6_36

DO - 10.1007/978-3-319-09858-6_36

M3 - Contribución a la conferencia

AN - SCOPUS:84928238742

T3 - Mechanisms and Machine Science

SP - 381

EP - 389

BT - Multibody Mechatronic Systems - Proceedings of the MUSME Conference, 2014

A2 - Martinez, Eusebio Eduardo Hernández

A2 - Ceccarelli, Marco

PB - Kluwer Academic Publishers

T2 - 5th International Symposium on Multibody Systems and Mechatronics, MUSME 2014

Y2 - 21 October 2014 through 24 October 2014

ER -

Identification of a cylindrical robot using recurrent neural networks

Resumen

Serie de la publicación

Conferencia

Acceder al documento

Otros archivos y enlaces

Huella

Citar esto